Free directory · 103 models · best benchmarks first

Every free uncensored AI model on the internet

Run them free on your own computer. Your strategies never leave your machine. Pick your computer below to see what runs on it, then copy one command to start. No GPU? Most have an API too.

Turn a model into a trading bot.AgenKit builds one from your idea with one prompt, in the AI coding tool you already use.Build it with AgenKit

103 models, junk and copies removed
18 run on an 8 GB laptop
52 ready to use by API
Scores from the UGI Leaderboard, Oct 2026

1What computer do you have?

PC graphics cardMac

How do I check?How to check

Windows: Task Manager → Performance → GPU → “Dedicated GPU memory”. Mac: Apple menu → About This Mac → “Memory”. A laptop without a gaming graphics card: pick “No GPU”.

2What do you want it for?

Sort

Showing all 103

gemma-4-31B-it-heretic
by coder3101 · built on Gemma 4 31B
31BThinkingVisionAPIApr 2026
Uncensored IQ
65.7
Reasoning
36.5
Answers
10/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots. Works the maths out step by step.
Needs about 21 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-4-31B-it-heretic-GGUF:Q4_K_M
Hugging Face
Use gemma-4-31B-it-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 18.7 GB once, then chat:
ollama run hf.co/mradermacher/gemma-4-31B-it-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-4-31B-it-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="coder3101/gemma-4-31B-it-heretic", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
gemma-4-31B-it-uncensored-heretic
by llmfan46 · built on Gemma 4 31B
31BThinkingVisionAPIApr 2026
Uncensored IQ
65.0
Reasoning
36.4
Answers
10/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots. Works the maths out step by step.
Needs about 21 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-4-31B-it-uncensored-heretic-GGUF:Q4_K_M
Hugging Face
Use gemma-4-31B-it-uncensored-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 18.7 GB once, then chat:
ollama run hf.co/mradermacher/gemma-4-31B-it-uncensored-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-4-31B-it-uncensored-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="llmfan46/gemma-4-31B-it-uncensored-heretic", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Assistant_Pepe_70B
by SicariusSicariiStuff · built on Llama 3.1 70B
70BAPIApr 2026
Uncensored IQ
59.5
Reasoning
35.3
Answers
9.5/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Assistant_Pepe_70B-GGUF:Q4_K_M
Hugging Face
Use Assistant_Pepe_70B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/Assistant_Pepe_70B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Assistant_Pepe_70B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="SicariusSicariiStuff/Assistant_Pepe_70B", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Golddiamondgold-Paperbliteration-L33-70b
by KaraKaraWitch · built on GoldDiamondGold Llama 3.3 70B
70BFeb 2026
Uncensored IQ
58.9
Reasoning
35.3
Answers
8.8/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Golddiamondgold-Paperbliteration-L33-70b-GGUF:Q4_K_M
Hugging Face
Use Golddiamondgold-Paperbliteration-L33-70b
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/Golddiamondgold-Paperbliteration-L33-70b-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Golddiamondgold-Paperbliteration-L33-70b-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 44 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Behemoth-X-123B-v2
by TheDrummer · built on Mistral Large
123BAug 2025
Uncensored IQ
57.2
Reasoning
34.7
Answers
8.2/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Needs a big machine or an API.
Needs about 47 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Behemoth-X-123B-v2-GGUF:Q2_K
Hugging Face
Use Behemoth-X-123B-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 45.2 GB once, then chat:
ollama run hf.co/mradermacher/Behemoth-X-123B-v2-GGUF:Q2_K
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Behemoth-X-123B-v2-GGUF:Q2_K", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 47 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Qwen3-VL-235B-A22B-Thinking-heretic
by scrunter
235B · 22B activeThinkingVisionJan 2026
Uncensored IQ
58.1
Reasoning
33.7
Answers
9.2/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Reads chart screenshots. Works the maths out step by step.
Needs about 144 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3-VL-235B-A22B-Thinking-heretic-GGUF:Q4_K_M
Hugging Face
Use Qwen3-VL-235B-A22B-Thinking-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 142.2 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3-VL-235B-A22B-Thinking-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3-VL-235B-A22B-Thinking-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 144 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Qwen3.5-Queen-27B
by aifeifei798 · built on Qwen3.5 27B
27BThinkingVisionMar 2026
Uncensored IQ
57.7
Reasoning
33.6
Answers
10/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots. Works the maths out step by step.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3.5-Queen-27B-GGUF:Q4_K_M
Hugging Face
Use Qwen3.5-Queen-27B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.5 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3.5-Queen-27B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3.5-Queen-27B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 18 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
GLM-4.5-Air-Derestricted
by ArliAI · built on GLM 4.5 Air
106B · 12B activeThinkingNov 2025
Uncensored IQ
53.3
Reasoning
36.6
Answers
8.8/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Works the maths out step by step. Fast enough to run inside a bot all day.
Needs about 47 GB of graphics or Mac memory
ollama run hf.co/mradermacher/GLM-4.5-Air-Derestricted-GGUF:Q2_K
Hugging Face
Use GLM-4.5-Air-Derestricted
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 45 GB once, then chat:
ollama run hf.co/mradermacher/GLM-4.5-Air-Derestricted-GGUF:Q2_K
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/GLM-4.5-Air-Derestricted-GGUF:Q2_K", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 47 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gemma-4-26B-A4B-it-SOMPOA-heresy
by MuXodious · built on Gemma 4 26B A4B
26B · 4B activeThinkingVisionAPIApr 2026
Uncensored IQ
54.0
Reasoning
34.5
Answers
9.5/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots. Works the maths out step by step.
Needs about 19 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-4-26B-A4B-it-SOMPOA-heresy-GGUF:Q4_K_M
Hugging Face
Use gemma-4-26B-A4B-it-SOMPOA-heresy
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.8 GB once, then chat:
ollama run hf.co/mradermacher/gemma-4-26B-A4B-it-SOMPOA-heresy-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-4-26B-A4B-it-SOMPOA-heresy-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="MuXodious/gemma-4-26B-A4B-it-SOMPOA-heresy", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
gemma-4-26B-A4B-it-ultra-uncensored-heretic
by llmfan46 · built on Gemma 4 26B A4B
26B · 4B activeThinkingVisionAPIApr 2026
Uncensored IQ
52.4
Reasoning
35.7
Answers
10/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots. Works the maths out step by step.
Needs about 19 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-4-26B-A4B-it-ultra-uncensored-heretic-GGUF:Q4_K_M
Hugging Face
Use gemma-4-26B-A4B-it-ultra-uncensored-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.8 GB once, then chat:
ollama run hf.co/mradermacher/gemma-4-26B-A4B-it-ultra-uncensored-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-4-26B-A4B-it-ultra-uncensored-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="llmfan46/gemma-4-26B-A4B-it-ultra-uncensored-heretic", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Qwen3.5-27B-heretic
by coder3101 · built on Qwen3.5 27B
27BThinkingVisionFeb 2026
Uncensored IQ
54.2
Reasoning
33.8
Answers
10/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots. Works the maths out step by step.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3.5-27B-heretic-GGUF:Q4_K_M
Hugging Face
Use Qwen3.5-27B-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.5 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3.5-27B-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3.5-27B-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 18 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Sapphira-L3.3-70b-0.1
by BruhzWater
70BAPIJul 2025
Uncensored IQ
56.0
Reasoning
30.9
Answers
8/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Sapphira-L3.3-70b-0.1-GGUF:Q4_K_M
Hugging Face
Use Sapphira-L3.3-70b-0.1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/Sapphira-L3.3-70b-0.1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Sapphira-L3.3-70b-0.1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="BruhzWater/Sapphira-L3.3-70b-0.1", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Precog-123B-v1
by TheDrummer · built on Mistral Large
123BThinkingNov 2025
Uncensored IQ
52.5
Reasoning
33.8
Answers
7/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Works the maths out step by step. Needs a big machine or an API.
Needs about 47 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Precog-123B-v1-GGUF:Q2_K
Hugging Face
Use Precog-123B-v1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 45.2 GB once, then chat:
ollama run hf.co/mradermacher/Precog-123B-v1-GGUF:Q2_K
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Precog-123B-v1-GGUF:Q2_K", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 47 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gpt-oss-120b-tainted-heresy
by MuXodious · built on gpt oss 120B
120B · 5.1B activeThinkingFeb 2026
Uncensored IQ
53.8
Reasoning
32.3
Answers
9.2/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Works the maths out step by step. Fast enough to run inside a bot all day.
Needs about 90 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gpt-oss-120b-tainted-heresy-GGUF:Q4_K_M
Hugging Face
Use gpt-oss-120b-tainted-heresy
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 87.9 GB once, then chat:
ollama run hf.co/mradermacher/gpt-oss-120b-tainted-heresy-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gpt-oss-120b-tainted-heresy-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 90 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Stylizer-V2-LLaMa-70B
by TareksGraveyard · built on Llama 3.1 Nemotron lorablated 70B
70BAPIApr 2025
Uncensored IQ
56.7
Reasoning
28.8
Answers
7.8/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Stylizer-V2-LLaMa-70B-GGUF:Q4_K_M
Hugging Face
Use Stylizer-V2-LLaMa-70B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/Stylizer-V2-LLaMa-70B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Stylizer-V2-LLaMa-70B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="TareksGraveyard/Stylizer-V2-LLaMa-70B", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Qwen3.5-35B-A3B-uncensored-heretic
by llmfan46 · built on Qwen3.5 35B A3B
35B · 3B activeThinkingVisionFeb 2026
Uncensored IQ
49.6
Reasoning
35.6
Answers
10/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots. Works the maths out step by step.
Needs about 23 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3.5-35B-A3B-uncensored-heretic-GGUF:Q4_K_M
Hugging Face
Use Qwen3.5-35B-A3B-uncensored-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 21.2 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3.5-35B-A3B-uncensored-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3.5-35B-A3B-uncensored-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 23 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gpt-4o-distil-Llama-3.3-70B-Instruct-PaperWitch-heresy
by MuXodious · built on gpt 4o distil Llama 3.3 70B
70BAPIFeb 2026
Uncensored IQ
54.3
Reasoning
30.6
Answers
8.8/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gpt-4o-distil-Llama-3.3-70B-Instruct-PaperWitch-heresy-GGUF:Q4_K_M
Hugging Face
Use gpt-4o-distil-Llama-3.3-70B-Instruct-PaperWitch-heresy
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/gpt-4o-distil-Llama-3.3-70B-Instruct-PaperWitch-heresy-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gpt-4o-distil-Llama-3.3-70B-Instruct-PaperWitch-heresy-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="MuXodious/gpt-4o-distil-Llama-3.3-70B-Instruct-PaperWitch-heresy", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Qwen3.5-35B-A3B-heretic-v2
by llmfan46 · built on Qwen3.5 35B A3B
35B · 3B activeThinkingVisionMar 2026
Uncensored IQ
49.6
Reasoning
34.3
Answers
10/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots. Works the maths out step by step.
Needs about 23 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3.5-35B-A3B-heretic-v2-GGUF:Q4_K_M
Hugging Face
Use Qwen3.5-35B-A3B-heretic-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 21.2 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3.5-35B-A3B-heretic-v2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3.5-35B-A3B-heretic-v2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 23 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Anubis-70B-v1.1
by TheDrummer · built on Llama 3.3 70B
70BAPIJun 2025
Uncensored IQ
50.8
Reasoning
32.5
Answers
8.5/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Anubis-70B-v1.1-GGUF:Q4_K_M
Hugging Face
Use Anubis-70B-v1.1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/Anubis-70B-v1.1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Anubis-70B-v1.1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="TheDrummer/Anubis-70B-v1.1", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Behemoth-X-123B-v2.1
by TheDrummer
123BOct 2025
Uncensored IQ
49.9
Reasoning
32.8
Answers
8.2/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Needs a big machine or an API.
Needs about 75 GB of graphics or Mac memory
llama-server -hf TheDrummer/Behemoth-X-123B-v2.1-GGUF:Q4_K_M
Hugging Face
Use Behemoth-X-123B-v2.1
Install llama.cpp (free). This model is split into parts, which llama.cpp downloads for you.
Paste this in Terminal. It downloads about 73.2 GB once, then chat:
llama-server -hf TheDrummer/Behemoth-X-123B-v2.1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:8080/v1", api_key="local") r = ai.chat.completions.create(model="TheDrummer/Behemoth-X-123B-v2.1-GGUF", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 75 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Heretic-Qwen3-VL-235B-A22B-Instruct
by KaraKaraWitch · built on Qwen3 VL 235B A22B
235B · 22B activeVisionDec 2025
Uncensored IQ
54.6
Reasoning
28.0
Answers
8.5/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Reads chart screenshots. Needs a big machine or an API.
Needs about 144 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Heretic-Qwen3-VL-235B-A22B-Instruct-GGUF:Q4_K_M
Hugging Face
Use Heretic-Qwen3-VL-235B-A22B-Instruct
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 142.2 GB once, then chat:
ollama run hf.co/mradermacher/Heretic-Qwen3-VL-235B-A22B-Instruct-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Heretic-Qwen3-VL-235B-A22B-Instruct-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 144 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Qwen3.6-35B-A3B-uncensored-heretic
by llmfan46 · built on Qwen3.6 35B A3B
35B · 3B activeThinkingVisionApr 2026
Uncensored IQ
50.9
Reasoning
31.6
Answers
10/10
For trading: Solid market research and plain-English trade plans. Reads chart screenshots. Works the maths out step by step.
Needs about 23 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3.6-35B-A3B-uncensored-heretic-GGUF:Q4_K_M
Hugging Face
Use Qwen3.6-35B-A3B-uncensored-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 21.2 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3.6-35B-A3B-uncensored-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3.6-35B-A3B-uncensored-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 23 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
GoldDiamondGold-Abliterated-L33-70b
by KaraKaraWitch · built on GoldDiamondGold Llama 3.3 70B
70BAPIFeb 2026
Uncensored IQ
51.7
Reasoning
30.6
Answers
8/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/GoldDiamondGold-Abliterated-L33-70b-GGUF:Q4_K_M
Hugging Face
Use GoldDiamondGold-Abliterated-L33-70b
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/GoldDiamondGold-Abliterated-L33-70b-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/GoldDiamondGold-Abliterated-L33-70b-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="KaraKaraWitch/GoldDiamondGold-Abliterated-L33-70b", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Llama-3.3-70B-Instruct-abliterated
by huihui-ai · built on Llama 3.3 70B
70BAPIDec 2024
Uncensored IQ
48.4
Reasoning
33.5
Answers
8/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/bartowski/Llama-3.3-70B-Instruct-abliterated-GGUF:Q4_K_M
Hugging Face
Use Llama-3.3-70B-Instruct-abliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/bartowski/Llama-3.3-70B-Instruct-abliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/bartowski/Llama-3.3-70B-Instruct-abliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="huihui-ai/Llama-3.3-70B-Instruct-abliterated", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Citrus1.0-llama-70B
by jdh-algo
70BAPIMar 2025
Uncensored IQ
48.0
Reasoning
33.7
Answers
7/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Citrus1.0-llama-70B-GGUF:Q4_K_M
Hugging Face
Use Citrus1.0-llama-70B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/Citrus1.0-llama-70B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Citrus1.0-llama-70B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="jdh-algo/Citrus1.0-llama-70B", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
gpt4o-distil-paperwitch-abliteration-L33-70b
by KaraKaraWitch · built on gpt 4o distil Llama 3.3 70B
70BFeb 2026
Uncensored IQ
51.0
Reasoning
30.7
Answers
8/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gpt4o-distil-paperwitch-abliteration-L33-70b-GGUF:Q4_K_M
Hugging Face
Use gpt4o-distil-paperwitch-abliteration-L33-70b
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/gpt4o-distil-paperwitch-abliteration-L33-70b-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gpt4o-distil-paperwitch-abliteration-L33-70b-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 44 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gpt-oss-120b-Derestricted
by ArliAI · built on gpt oss 120B
120B · 5.1B activeThinkingNov 2025
Uncensored IQ
45.1
Reasoning
35.7
Answers
9.2/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Works the maths out step by step. Fast enough to run inside a bot all day.
Needs about 90 GB of graphics or Mac memory
ollama run hf.co/Calandracas/gpt-oss-120b-Derestricted-GGUF:Q4_K_M
Hugging Face
Use gpt-oss-120b-Derestricted
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 88.4 GB once, then chat:
ollama run hf.co/Calandracas/gpt-oss-120b-Derestricted-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/Calandracas/gpt-oss-120b-Derestricted-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 90 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Golem-70B-v1a
by Naphula
70BNov 2025
Uncensored IQ
48.1
Reasoning
32.7
Answers
7.2/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Golem-70B-v1a-GGUF:Q4_K_M
Hugging Face
Use Golem-70B-v1a
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/Golem-70B-v1a-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Golem-70B-v1a-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 44 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Rotor_24B_V.1
by OddTheGreat · built on MS3.2 PaintedFantasy v2 24B
24BOct 2025
Uncensored IQ
50.2
Reasoning
29.4
Answers
8.5/10
For trading: Solid market research and plain-English trade plans.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Rotor_24B_V.1-GGUF:Q4_K_M
Hugging Face
Use Rotor_24B_V.1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Rotor_24B_V.1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Rotor_24B_V.1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gemma-3-27b-it-abliterated-normpreserve
by YanLabs · built on Gemma 3 27B
27BVisionDec 2025
Uncensored IQ
45.0
Reasoning
34.1
Answers
9.5/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-3-27b-it-abliterated-normpreserve-GGUF:Q4_K_M
Hugging Face
Use gemma-3-27b-it-abliterated-normpreserve
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.5 GB once, then chat:
ollama run hf.co/mradermacher/gemma-3-27b-it-abliterated-normpreserve-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-3-27b-it-abliterated-normpreserve-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 18 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
WeirdCompound-v1.7-24b
by FlareRebellion
24BAPIOct 2025
Uncensored IQ
48.0
Reasoning
30.2
Answers
8/10
For trading: Solid market research and plain-English trade plans.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/WeirdCompound-v1.7-24b-GGUF:Q4_K_M
Hugging Face
Use WeirdCompound-v1.7-24b
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/WeirdCompound-v1.7-24b-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/WeirdCompound-v1.7-24b-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="FlareRebellion/WeirdCompound-v1.7-24b", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Llama-3.1-Nemotron-lorablated-70B
by nbeerbower · built on Llama 3.1 Nemotron 70B
70BAPIOct 2024
Uncensored IQ
47.4
Reasoning
30.3
Answers
7.5/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Llama-3.1-Nemotron-lorablated-70B-GGUF:Q4_K_M
Hugging Face
Use Llama-3.1-Nemotron-lorablated-70B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/Llama-3.1-Nemotron-lorablated-70B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Llama-3.1-Nemotron-lorablated-70B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="nbeerbower/Llama-3.1-Nemotron-lorablated-70B", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Goetia-26B-A4B-v1.3-Absolute-Heretic-ARA
by Naphula · built on Gemma 4 26B A4B
26B · 4B activeThinkingVisionAPIJun 2026
Uncensored IQ
44.1
Reasoning
33.5
Answers
9.5/10
For trading: Sharp reasoning for strategy ideas, trade plans and backtest reviews. Reads chart screenshots. Works the maths out step by step.
Needs about 19 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Goetia-26B-A4B-v1.3-Absolute-Heretic-ARA-GGUF:Q4_K_M
Hugging Face
Use Goetia-26B-A4B-v1.3-Absolute-Heretic-ARA
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.8 GB once, then chat:
ollama run hf.co/mradermacher/Goetia-26B-A4B-v1.3-Absolute-Heretic-ARA-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Goetia-26B-A4B-v1.3-Absolute-Heretic-ARA-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="Naphula/Goetia-26B-A4B-v1.3-Absolute-Heretic-ARA", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Agatha-111B-v1
by TheDrummer · built on Command A
111BJun 2025
Uncensored IQ
45.3
Reasoning
32.0
Answers
7.2/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Needs a big machine or an API.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Agatha-111B-v1-GGUF:Q2_K
Hugging Face
Use Agatha-111B-v1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.1 GB once, then chat:
ollama run hf.co/mradermacher/Agatha-111B-v1-GGUF:Q2_K
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Agatha-111B-v1-GGUF:Q2_K", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 44 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
L3-70B-Euryale-v2.1
by Sao10K
70BAPIJun 2024
Uncensored IQ
49.3
Reasoning
27.6
Answers
7.5/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/bartowski/L3-70B-Euryale-v2.1-GGUF:Q4_K_M
Hugging Face
Use L3-70B-Euryale-v2.1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/bartowski/L3-70B-Euryale-v2.1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/bartowski/L3-70B-Euryale-v2.1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="Sao10K/L3-70B-Euryale-v2.1", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
BereavedCompound-v1.0-24b
by FlareRebellion
24BAPINov 2025
Uncensored IQ
48.7
Reasoning
28.1
Answers
8/10
For trading: Solid market research and plain-English trade plans.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/BereavedCompound-v1.0-24b-GGUF:Q4_K_M
Hugging Face
Use BereavedCompound-v1.0-24b
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/BereavedCompound-v1.0-24b-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/BereavedCompound-v1.0-24b-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="FlareRebellion/BereavedCompound-v1.0-24b", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Llama-3.1-70B-ArliAI-RPMax-v1.1
by ArliAI
70BAPISep 2024
Uncensored IQ
50.0
Reasoning
26.7
Answers
7.5/10
For trading: Quick, direct answers and market summaries.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/bartowski/Llama-3.1-70B-ArliAI-RPMax-v1.1-GGUF:Q4_K_M
Hugging Face
Use Llama-3.1-70B-ArliAI-RPMax-v1.1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/bartowski/Llama-3.1-70B-ArliAI-RPMax-v1.1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/bartowski/Llama-3.1-70B-ArliAI-RPMax-v1.1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="ArliAI/Llama-3.1-70B-ArliAI-RPMax-v1.1", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
gemma-3-27b-it-ultra-uncensored-heretic
by llmfan46 · built on Gemma 3 27B
27BVisionMar 2026
Uncensored IQ
44.9
Reasoning
31.0
Answers
10/10
For trading: Solid market research and plain-English trade plans. Reads chart screenshots.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-3-27b-it-ultra-uncensored-heretic-GGUF:Q4_K_M
Hugging Face
Use gemma-3-27b-it-ultra-uncensored-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.5 GB once, then chat:
ollama run hf.co/mradermacher/gemma-3-27b-it-ultra-uncensored-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-3-27b-it-ultra-uncensored-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 18 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
GLM-4-32B-0414-uncensored-heretic-v1
by llmfan46 · built on GLM 4 32B
32BAPIMar 2026
Uncensored IQ
50.5
Reasoning
25.0
Answers
10/10
For trading: Quick, direct answers and market summaries.
Needs about 22 GB of graphics or Mac memory
ollama run hf.co/mradermacher/GLM-4-32B-0414-uncensored-heretic-v1-GGUF:Q4_K_M
Hugging Face
Use GLM-4-32B-0414-uncensored-heretic-v1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 19.7 GB once, then chat:
ollama run hf.co/mradermacher/GLM-4-32B-0414-uncensored-heretic-v1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/GLM-4-32B-0414-uncensored-heretic-v1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="llmfan46/GLM-4-32B-0414-uncensored-heretic-v1", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Cydonia-24B-v4
by TheDrummer · built on Mistral Small 3.2 24B
24BAPIJul 2025
Uncensored IQ
48.2
Reasoning
26.6
Answers
7.8/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Cydonia-24B-v4-GGUF:Q4_K_M
Hugging Face
Use Cydonia-24B-v4
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Cydonia-24B-v4-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Cydonia-24B-v4-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="TheDrummer/Cydonia-24B-v4", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
MS3.2-Austral-Winton
by Delta-Vector · built on MS3.2 Austral 24B KTO
24BAPIJul 2025
Uncensored IQ
48.2
Reasoning
26.4
Answers
8/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/MS3.2-Austral-Winton-GGUF:Q4_K_M
Hugging Face
Use MS3.2-Austral-Winton
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/MS3.2-Austral-Winton-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/MS3.2-Austral-Winton-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="Delta-Vector/MS3.2-Austral-Winton", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Skyfall-31B-v4
by TheDrummer · built on Mistral Small 3.2 24B
31BAug 2025
Uncensored IQ
50.7
Reasoning
23.9
Answers
8/10
For trading: Quick, direct answers and market summaries.
Needs about 21 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Skyfall-31B-v4-GGUF:Q4_K_M
Hugging Face
Use Skyfall-31B-v4
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 19 GB once, then chat:
ollama run hf.co/mradermacher/Skyfall-31B-v4-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Skyfall-31B-v4-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 21 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Mechanism_24B_V.1
by OddTheGreat · built on Codex 24B Small 3.2
24BJul 2025
Uncensored IQ
48.4
Reasoning
25.6
Answers
7/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Mechanism_24B_V.1-GGUF:Q4_K_M
Hugging Face
Use Mechanism_24B_V.1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Mechanism_24B_V.1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Mechanism_24B_V.1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Seed-OSS-36B-Instruct-biprojected-norm-preserving-abliterated
by vprilepskii · built on Seed OSS 36B
36BThinkingNov 2025
Uncensored IQ
46.1
Reasoning
27.9
Answers
8.5/10
For trading: Solid market research and plain-English trade plans. Works the maths out step by step.
Needs about 24 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Seed-OSS-36B-Instruct-biprojected-norm-preserving-abliterated-GGUF:Q4_K_M
Hugging Face
Use Seed-OSS-36B-Instruct-biprojected-norm-preserving-abliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 21.8 GB once, then chat:
ollama run hf.co/mradermacher/Seed-OSS-36B-Instruct-biprojected-norm-preserving-abliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Seed-OSS-36B-Instruct-biprojected-norm-preserving-abliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 24 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Huihui-Qwen3-VL-32B-Thinking-abliterated
by huihui-ai · built on Qwen3 VL 32B Thinking
32BThinkingVisionAPIOct 2025
Uncensored IQ
50.4
Reasoning
22.8
Answers
10/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Works the maths out step by step.
Needs about 22 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Huihui-Qwen3-VL-32B-Thinking-abliterated-GGUF:Q4_K_M
Hugging Face
Use Huihui-Qwen3-VL-32B-Thinking-abliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 19.8 GB once, then chat:
ollama run hf.co/mradermacher/Huihui-Qwen3-VL-32B-Thinking-abliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Huihui-Qwen3-VL-32B-Thinking-abliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="huihui-ai/Huihui-Qwen3-VL-32B-Thinking-abliterated", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
MS3.2-24B-Penumbra-Aether
by Vortex5 · built on Hearthfire 24B
24BJan 2026
Uncensored IQ
46.5
Reasoning
26.7
Answers
7.8/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/MS3.2-24B-Penumbra-Aether-GGUF:Q4_K_M
Hugging Face
Use MS3.2-24B-Penumbra-Aether
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/MS3.2-24B-Penumbra-Aether-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/MS3.2-24B-Penumbra-Aether-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Big-Tiger-Gemma-27B-v3-heretic-v2
by coder3101 · built on Big Tiger Gemma 27B v3
27BVisionAPIDec 2025
Uncensored IQ
44.6
Reasoning
28.4
Answers
8.5/10
For trading: Solid market research and plain-English trade plans. Reads chart screenshots.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Big-Tiger-Gemma-27B-v3-heretic-v2-GGUF:Q4_K_M
Hugging Face
Use Big-Tiger-Gemma-27B-v3-heretic-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.5 GB once, then chat:
ollama run hf.co/mradermacher/Big-Tiger-Gemma-27B-v3-heretic-v2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Big-Tiger-Gemma-27B-v3-heretic-v2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="coder3101/Big-Tiger-Gemma-27B-v3-heretic-v2", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
GLM-4.5-Iceblink-v2-106B-A12B
by zerofata · built on GLM 4.5 Air
106B · 12B activeNov 2025
Uncensored IQ
48.0
Reasoning
24.8
Answers
7.2/10
For trading: The deepest analysis here: filings, macro and full strategy write-ups. Fast enough to run inside a bot all day.
Needs about 47 GB of graphics or Mac memory
ollama run hf.co/mradermacher/GLM-4.5-Iceblink-v2-106B-A12B-GGUF:Q2_K
Hugging Face
Use GLM-4.5-Iceblink-v2-106B-A12B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 45 GB once, then chat:
ollama run hf.co/mradermacher/GLM-4.5-Iceblink-v2-106B-A12B-GGUF:Q2_K
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/GLM-4.5-Iceblink-v2-106B-A12B-GGUF:Q2_K", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 47 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Goetia-24B-v1
by Naphula · built on Eurydice 24B v3.5
24BOct 2025
Uncensored IQ
47.7
Reasoning
25.0
Answers
7.5/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Goetia-24B-v1-GGUF:Q4_K_M
Hugging Face
Use Goetia-24B-v1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Goetia-24B-v1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Goetia-24B-v1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Harbinger-24B-absolute-heresy
by MuXodious · built on Harbinger 24B
24BJan 2026
Uncensored IQ
44.3
Reasoning
28.2
Answers
8.2/10
For trading: Solid market research and plain-English trade plans.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Harbinger-24B-absolute-heresy-GGUF:Q4_K_M
Hugging Face
Use Harbinger-24B-absolute-heresy
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Harbinger-24B-absolute-heresy-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Harbinger-24B-absolute-heresy-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
BlackSheep-24B
by TroyDoesAI
24BAPIMar 2025
Uncensored IQ
48.1
Reasoning
24.2
Answers
9.5/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/BlackSheep-24B-GGUF:Q4_K_M
Hugging Face
Use BlackSheep-24B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/BlackSheep-24B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/BlackSheep-24B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="TroyDoesAI/BlackSheep-24B", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Qwen3-VL-32B-Thinking-heretic-v2
by coder3101 · built on Qwen3 VL 32B Thinking
32BThinkingVisionAPIDec 2025
Uncensored IQ
47.5
Reasoning
24.7
Answers
9.2/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Works the maths out step by step.
Needs about 22 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3-VL-32B-Thinking-heretic-v2-GGUF:Q4_K_M
Hugging Face
Use Qwen3-VL-32B-Thinking-heretic-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 19.8 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3-VL-32B-Thinking-heretic-v2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3-VL-32B-Thinking-heretic-v2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="coder3101/Qwen3-VL-32B-Thinking-heretic-v2", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
gemma3-27b-abliterated-dpo
by summykai · built on Gemma 3 27B abliterated
27BVisionAPIApr 2025
Uncensored IQ
41.8
Reasoning
29.8
Answers
8.8/10
For trading: Solid market research and plain-English trade plans. Reads chart screenshots.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma3-27b-abliterated-dpo-GGUF:Q4_K_M
Hugging Face
Use gemma3-27b-abliterated-dpo
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.5 GB once, then chat:
ollama run hf.co/mradermacher/gemma3-27b-abliterated-dpo-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma3-27b-abliterated-dpo-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="summykai/gemma3-27b-abliterated-dpo", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Circuitry_24B_V.2
by OddTheGreat · built on Rei 24B KTO
24BAPISep 2025
Uncensored IQ
49.0
Reasoning
22.5
Answers
7.8/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Circuitry_24B_V.2-GGUF:Q4_K_M
Hugging Face
Use Circuitry_24B_V.2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Circuitry_24B_V.2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Circuitry_24B_V.2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="OddTheGreat/Circuitry_24B_V.2", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Codex-24B-Small-3.2
by Gryphe · built on Mistral Small 3.2 24B
24BAPIJun 2025
Uncensored IQ
45.8
Reasoning
25.4
Answers
8/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Codex-24B-Small-3.2-GGUF:Q4_K_M
Hugging Face
Use Codex-24B-Small-3.2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Codex-24B-Small-3.2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Codex-24B-Small-3.2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="Gryphe/Codex-24B-Small-3.2", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
M3.2-24B-Loki-V1.3
by CrucibleLab
24BAPIAug 2025
Uncensored IQ
42.4
Reasoning
28.8
Answers
7/10
For trading: Solid market research and plain-English trade plans.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/M3.2-24B-Loki-V1.3-GGUF:Q4_K_M
Hugging Face
Use M3.2-24B-Loki-V1.3
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/M3.2-24B-Loki-V1.3-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/M3.2-24B-Loki-V1.3-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="CrucibleLab/M3.2-24B-Loki-V1.3", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
MiniusLight-24B-v3
by DoppelReflEx · built on Rei 24B KTO
24BAug 2025
Uncensored IQ
47.1
Reasoning
24.0
Answers
7.5/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/MiniusLight-24B-v3-GGUF:Q4_K_M
Hugging Face
Use MiniusLight-24B-v3
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/MiniusLight-24B-v3-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/MiniusLight-24B-v3-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Hearthfire-24B-absolute-heresy
by MuXodious · built on Hearthfire 24B
24BAPIJan 2026
Uncensored IQ
48.1
Reasoning
22.8
Answers
9/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Hearthfire-24B-absolute-heresy-GGUF:Q4_K_M
Hugging Face
Use Hearthfire-24B-absolute-heresy
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Hearthfire-24B-absolute-heresy-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Hearthfire-24B-absolute-heresy-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="MuXodious/Hearthfire-24B-absolute-heresy", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Q3.5-BlueStar-27B-ultra-heretic
by llmfan46 · built on Q3.5 BlueStar 27B
27BVisionAPIMar 2026
Uncensored IQ
47.5
Reasoning
22.9
Answers
9/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Q3.5-BlueStar-27B-ultra-heretic-GGUF:Q4_K_M
Hugging Face
Use Q3.5-BlueStar-27B-ultra-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.5 GB once, then chat:
ollama run hf.co/mradermacher/Q3.5-BlueStar-27B-ultra-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Q3.5-BlueStar-27B-ultra-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="llmfan46/Q3.5-BlueStar-27B-ultra-heretic", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Qwen3-VL-32B-Instruct-heretic-v2
by coder3101 · built on Qwen3 VL 32B
32BVisionAPIDec 2025
Uncensored IQ
45.3
Reasoning
25.1
Answers
9.5/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots.
Needs about 22 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3-VL-32B-Instruct-heretic-v2-GGUF:Q4_K_M
Hugging Face
Use Qwen3-VL-32B-Instruct-heretic-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 19.8 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3-VL-32B-Instruct-heretic-v2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3-VL-32B-Instruct-heretic-v2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="coder3101/Qwen3-VL-32B-Instruct-heretic-v2", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved
by llmfan46 · built on Qwen3.6 27B
27BThinkingVisionAPIMay 2026
Uncensored IQ
43.9
Reasoning
26.3
Answers
9.5/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Works the maths out step by step.
Needs about 19 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved-GGUF:Q4_K_M
Hugging Face
Use Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.8 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="llmfan46/Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Cydonia-24B-v4.3-heretic
by coder3101 · built on Cydonia 24B v4.3
24BThinkingAPIDec 2025
Uncensored IQ
48.2
Reasoning
21.9
Answers
10/10
For trading: Quick, direct answers and market summaries. Works the maths out step by step.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Cydonia-24B-v4.3-heretic-GGUF:Q4_K_M
Hugging Face
Use Cydonia-24B-v4.3-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Cydonia-24B-v4.3-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Cydonia-24B-v4.3-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="coder3101/Cydonia-24B-v4.3-heretic", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Boreas-24B-v1.2
by Naphula · built on Arcee Blitz
24BDec 2025
Uncensored IQ
43.5
Reasoning
26.4
Answers
7.2/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Boreas-24B-v1.2-GGUF:Q4_K_M
Hugging Face
Use Boreas-24B-v1.2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Boreas-24B-v1.2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Boreas-24B-v1.2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Sketch-Cydonia
by Ateron · built on Cydonia 24B v4.3
24BThinkingFeb 2026
Uncensored IQ
45.2
Reasoning
24.7
Answers
8/10
For trading: Quick, direct answers and market summaries. Works the maths out step by step.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Sketch-Cydonia-GGUF:Q4_K_M
Hugging Face
Use Sketch-Cydonia
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Sketch-Cydonia-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Sketch-Cydonia-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Harbinger-24B-biprojected-norm-preserving-abliterated
by vprilepskii · built on Harbinger 24B
24BNov 2025
Uncensored IQ
43.6
Reasoning
26.3
Answers
7.8/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Harbinger-24B-biprojected-norm-preserving-abliterated-GGUF:Q4_K_M
Hugging Face
Use Harbinger-24B-biprojected-norm-preserving-abliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Harbinger-24B-biprojected-norm-preserving-abliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Harbinger-24B-biprojected-norm-preserving-abliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gpt-oss-20b-RichardErkhov-heresy
by MuXodious · built on gpt oss 20B
20B · 3.6B activeThinkingFeb 2026
Uncensored IQ
46.7
Reasoning
22.5
Answers
10/10
For trading: Quick, direct answers and market summaries. Works the maths out step by step. Fast enough to run inside a bot all day.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gpt-oss-20b-RichardErkhov-heresy-GGUF:Q4_K_M
Hugging Face
Use gpt-oss-20b-RichardErkhov-heresy
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 15.8 GB once, then chat:
ollama run hf.co/mradermacher/gpt-oss-20b-RichardErkhov-heresy-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gpt-oss-20b-RichardErkhov-heresy-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 18 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
zetasepic-abliteratedV2-Qwen2.5-32B-Inst-BaseMerge-TIES
by CombinHorizon · built on Qwen2.5 32B
32BAPIDec 2024
Uncensored IQ
47.0
Reasoning
21.5
Answers
8.2/10
For trading: Quick, direct answers and market summaries.
Needs about 22 GB of graphics or Mac memory
ollama run hf.co/mradermacher/zetasepic-abliteratedV2-Qwen2.5-32B-Inst-BaseMerge-TIES-GGUF:Q4_K_M
Hugging Face
Use zetasepic-abliteratedV2-Qwen2.5-32B-Inst-BaseMerge-TIES
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 19.9 GB once, then chat:
ollama run hf.co/mradermacher/zetasepic-abliteratedV2-Qwen2.5-32B-Inst-BaseMerge-TIES-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/zetasepic-abliteratedV2-Qwen2.5-32B-Inst-BaseMerge-TIES-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="CombinHorizon/zetasepic-abliteratedV2-Qwen2.5-32B-Inst-BaseMerge-TIES", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
StormSeeker-24B-v1
by Naphula · built on M3.2 24B Loki V2
24BAPIJan 2026
Uncensored IQ
46.4
Reasoning
21.9
Answers
7.8/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/StormSeeker-24B-v1-GGUF:Q4_K_M
Hugging Face
Use StormSeeker-24B-v1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/StormSeeker-24B-v1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/StormSeeker-24B-v1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="Naphula/StormSeeker-24B-v1", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Harbinger-24B-noslop-absolute-heresy
by MuXodious · built on Harbinger 24B absolute heresy
24BJan 2026
Uncensored IQ
40.1
Reasoning
28.0
Answers
8.2/10
For trading: Solid market research and plain-English trade plans.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Harbinger-24B-noslop-absolute-heresy-GGUF:Q4_K_M
Hugging Face
Use Harbinger-24B-noslop-absolute-heresy
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Harbinger-24B-noslop-absolute-heresy-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Harbinger-24B-noslop-absolute-heresy-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Apparatus_24B
by OddTheGreat · built on Machina 24B.V2
24BFeb 2025
Uncensored IQ
41.2
Reasoning
26.8
Answers
7/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Apparatus_24B-GGUF:Q4_K_M
Hugging Face
Use Apparatus_24B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Apparatus_24B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Apparatus_24B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Cydonia-24B-v4.3
by TheDrummer · built on Mistral Small 3.2 24B
24BThinkingAPINov 2025
Uncensored IQ
45.7
Reasoning
21.8
Answers
7.8/10
For trading: Quick, direct answers and market summaries. Works the maths out step by step.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Cydonia-24B-v4.3-GGUF:Q4_K_M
Hugging Face
Use Cydonia-24B-v4.3
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Cydonia-24B-v4.3-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Cydonia-24B-v4.3-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="TheDrummer/Cydonia-24B-v4.3", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
L3.3-GeneticLemonade-Unleashed-v3-70B
by zerofata · built on Llama 3.3 GeneticLemonade Unleashed 70B
70BAPIMay 2025
Uncensored IQ
39.6
Reasoning
27.8
Answers
7.2/10
For trading: Solid market research and plain-English trade plans.
Needs about 44 GB of graphics or Mac memory
ollama run hf.co/mradermacher/L3.3-GeneticLemonade-Unleashed-v3-70B-GGUF:Q4_K_M
Hugging Face
Use L3.3-GeneticLemonade-Unleashed-v3-70B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 42.5 GB once, then chat:
ollama run hf.co/mradermacher/L3.3-GeneticLemonade-Unleashed-v3-70B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/L3.3-GeneticLemonade-Unleashed-v3-70B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="zerofata/L3.3-GeneticLemonade-Unleashed-v3-70B", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Magistral-Small-2509-ultra-uncensored-heretic-v1
by llmfan46 · built on Magistral Small
24BVisionAPIMar 2026
Uncensored IQ
43.5
Reasoning
23.8
Answers
8.5/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Magistral-Small-2509-ultra-uncensored-heretic-v1-GGUF:Q4_K_M
Hugging Face
Use Magistral-Small-2509-ultra-uncensored-heretic-v1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Magistral-Small-2509-ultra-uncensored-heretic-v1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Magistral-Small-2509-ultra-uncensored-heretic-v1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="llmfan46/Magistral-Small-2509-ultra-uncensored-heretic-v1", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Qwen3.5-4B-SOMPOA-heresy-v2
by MuXodious · built on Qwen3.5 4B
4BThinkingVisionMar 2026
Uncensored IQ
44.7
Reasoning
22.4
Answers
10/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Works the maths out step by step.
Needs about 5 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3.5-4B-SOMPOA-heresy-v2-GGUF:Q4_K_M
Hugging Face
Use Qwen3.5-4B-SOMPOA-heresy-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 2.7 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3.5-4B-SOMPOA-heresy-v2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3.5-4B-SOMPOA-heresy-v2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 5 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Qwen3.6-27B-uncensored-heretic-v2
by llmfan46 · built on Qwen3.6 27B
27BVisionAPIApr 2026
Uncensored IQ
43.9
Reasoning
22.9
Answers
9.5/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/llmfan46/Qwen3.6-27B-uncensored-heretic-v2-GGUF:Q4_K_M
Hugging Face
Use Qwen3.6-27B-uncensored-heretic-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.5 GB once, then chat:
ollama run hf.co/llmfan46/Qwen3.6-27B-uncensored-heretic-v2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/llmfan46/Qwen3.6-27B-uncensored-heretic-v2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="llmfan46/Qwen3.6-27B-uncensored-heretic-v2", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Qwen2.5-32B-Instruct-abliterated-v2
by zetasepic · built on Qwen2.5 32B
32BAPIOct 2024
Uncensored IQ
44.3
Reasoning
21.9
Answers
7.8/10
For trading: Quick, direct answers and market summaries.
Needs about 22 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen2.5-32B-Instruct-abliterated-v2-GGUF:Q4_K_M
Hugging Face
Use Qwen2.5-32B-Instruct-abliterated-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 19.9 GB once, then chat:
ollama run hf.co/mradermacher/Qwen2.5-32B-Instruct-abliterated-v2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen2.5-32B-Instruct-abliterated-v2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="zetasepic/Qwen2.5-32B-Instruct-abliterated-v2", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Apriel-1.6-15b-Thinker-Magic_alpha-decensored_MPOA
by MagicalAlchemist · built on Apriel 1.6 15B Thinker
15BThinkingVisionFeb 2026
Uncensored IQ
41.3
Reasoning
24.6
Answers
9.2/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Works the maths out step by step.
Needs about 11 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Apriel-1.6-15b-Thinker-Magic_alpha-decensored_MPOA-GGUF:Q4_K_M
Hugging Face
Use Apriel-1.6-15b-Thinker-Magic_alpha-decensored_MPOA
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 8.8 GB once, then chat:
ollama run hf.co/mradermacher/Apriel-1.6-15b-Thinker-Magic_alpha-decensored_MPOA-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Apriel-1.6-15b-Thinker-Magic_alpha-decensored_MPOA-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 11 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Big-Tiger-Gemma-27B-v3-heretic
by coder3101 · built on Big Tiger Gemma 27B v3
27BDec 2025
Uncensored IQ
40.5
Reasoning
25.2
Answers
8.5/10
For trading: Quick, direct answers and market summaries.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Big-Tiger-Gemma-27B-v3-heretic-GGUF:Q4_K_M
Hugging Face
Use Big-Tiger-Gemma-27B-v3-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 16.5 GB once, then chat:
ollama run hf.co/mradermacher/Big-Tiger-Gemma-27B-v3-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Big-Tiger-Gemma-27B-v3-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 18 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Dans-PersonalityEngine-V1.2.0-24b
by PocketDoc · built on Mistral Small 24B Base
24BAPIFeb 2025
Uncensored IQ
39.5
Reasoning
26.0
Answers
7.5/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Dans-PersonalityEngine-V1.2.0-24b-GGUF:Q4_K_M
Hugging Face
Use Dans-PersonalityEngine-V1.2.0-24b
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Dans-PersonalityEngine-V1.2.0-24b-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Dans-PersonalityEngine-V1.2.0-24b-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="PocketDoc/Dans-PersonalityEngine-V1.2.0-24b", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Dolphin-Mistral-24B-Venice-Edition
by dphn · built on Mistral Small 24B
24BAPIJun 2025
Uncensored IQ
40.8
Reasoning
24.4
Answers
7.8/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Dolphin-Mistral-24B-Venice-Edition-GGUF:Q4_K_M
Hugging Face
Use Dolphin-Mistral-24B-Venice-Edition
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Dolphin-Mistral-24B-Venice-Edition-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Dolphin-Mistral-24B-Venice-Edition-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Make a free account and key at OpenRouter. You pay per use, no subscription.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://openrouter.ai/api/v1", api_key="YOUR_OPENROUTER_KEY") r = ai.chat.completions.create(model="cognitivecomputations/dolphin-mistral-24b-venice-edition", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Huihui-GLM-4.7-Flash-abliterated
by huihui-ai · built on GLM 4.7 Flash
30B · 3B activeThinkingJan 2026
Uncensored IQ
41.2
Reasoning
23.6
Answers
9/10
For trading: Quick, direct answers and market summaries. Works the maths out step by step. Fast enough to run inside a bot all day.
Needs about 20 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Huihui-GLM-4.7-Flash-abliterated-GGUF:Q4_K_M
Hugging Face
Use Huihui-GLM-4.7-Flash-abliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 18.1 GB once, then chat:
ollama run hf.co/mradermacher/Huihui-GLM-4.7-Flash-abliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Huihui-GLM-4.7-Flash-abliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 20 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Qwen3.5-9B-heretic-v2
by trohrbaugh · built on Qwen3.5 9B Base
9BThinkingVisionMar 2026
Uncensored IQ
40.5
Reasoning
24.2
Answers
9.5/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Works the maths out step by step.
Needs about 8 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3.5-9B-heretic-v2-GGUF:Q4_K_M
Hugging Face
Use Qwen3.5-9B-heretic-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 5.6 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3.5-9B-heretic-v2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3.5-9B-heretic-v2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 8 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
UnslopSmall-22B-v1
by TheDrummer
22BOct 2024
Uncensored IQ
43.2
Reasoning
21.0
Answers
7.2/10
For trading: Quick, direct answers and market summaries.
Needs about 15 GB of graphics or Mac memory
ollama run hf.co/mradermacher/UnslopSmall-22B-v1-GGUF:Q4_K_M
Hugging Face
Use UnslopSmall-22B-v1
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 13.3 GB once, then chat:
ollama run hf.co/mradermacher/UnslopSmall-22B-v1-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/UnslopSmall-22B-v1-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 15 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
GLM-4.7-Flash-absolute-heresy
by MuXodious · built on GLM 4.7 Flash
30B · 3B activeThinkingFeb 2026
Uncensored IQ
43.3
Reasoning
20.6
Answers
9.8/10
For trading: Quick, direct answers and market summaries. Works the maths out step by step. Fast enough to run inside a bot all day.
Needs about 20 GB of graphics or Mac memory
ollama run hf.co/mradermacher/GLM-4.7-Flash-absolute-heresy-GGUF:Q4_K_M
Hugging Face
Use GLM-4.7-Flash-absolute-heresy
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 18.1 GB once, then chat:
ollama run hf.co/mradermacher/GLM-4.7-Flash-absolute-heresy-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/GLM-4.7-Flash-absolute-heresy-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 20 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Qwen3.5-4B-ARA-heresy-v2
by MuXodious · built on Qwen3.5 4B
4BThinkingVisionMar 2026
Uncensored IQ
40.5
Reasoning
22.6
Answers
9/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Works the maths out step by step.
Needs about 5 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Qwen3.5-4B-ARA-heresy-v2-GGUF:Q4_K_M
Hugging Face
Use Qwen3.5-4B-ARA-heresy-v2
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 2.7 GB once, then chat:
ollama run hf.co/mradermacher/Qwen3.5-4B-ARA-heresy-v2-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Qwen3.5-4B-ARA-heresy-v2-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 5 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Huihui-Qwen3.5-9B-abliterated
by huihui-ai · built on Qwen3.5 9B
9BThinkingVisionAPIMar 2026
Uncensored IQ
43.2
Reasoning
19.8
Answers
9.5/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Works the maths out step by step.
Needs about 8 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Huihui-Qwen3.5-9B-abliterated-GGUF:Q4_K_M
Hugging Face
Use Huihui-Qwen3.5-9B-abliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 5.6 GB once, then chat:
ollama run hf.co/mradermacher/Huihui-Qwen3.5-9B-abliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Huihui-Qwen3.5-9B-abliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="huihui-ai/Huihui-Qwen3.5-9B-abliterated", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Magidonia-24B-v4.3-heretic
by coder3101 · built on Magidonia 24B v4.3
24BDec 2025
Uncensored IQ
41.7
Reasoning
21.2
Answers
7.2/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Magidonia-24B-v4.3-heretic-GGUF:Q4_K_M
Hugging Face
Use Magidonia-24B-v4.3-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/Magidonia-24B-v4.3-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Magidonia-24B-v4.3-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 16 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
MS3.2-PaintedFantasy-v4.1-24B
by zerofata · built on Magistral Small
24BAPIFeb 2026
Uncensored IQ
42.1
Reasoning
20.3
Answers
7.5/10
For trading: Quick, direct answers and market summaries.
Needs about 16 GB of graphics or Mac memory
ollama run hf.co/mradermacher/MS3.2-PaintedFantasy-v4.1-24B-GGUF:Q4_K_M
Hugging Face
Use MS3.2-PaintedFantasy-v4.1-24B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 14.3 GB once, then chat:
ollama run hf.co/mradermacher/MS3.2-PaintedFantasy-v4.1-24B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/MS3.2-PaintedFantasy-v4.1-24B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="zerofata/MS3.2-PaintedFantasy-v4.1-24B", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
EtherealAurora-12B-Lorablated
by yamatazen · built on EtherealAurora 12B
12BAPIAug 2025
Uncensored IQ
38.8
Reasoning
23.5
Answers
7.5/10
For trading: Quick, direct answers and market summaries. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/EtherealAurora-12B-Lorablated-GGUF:Q4_K_M
Hugging Face
Use EtherealAurora-12B-Lorablated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.5 GB once, then chat:
ollama run hf.co/mradermacher/EtherealAurora-12B-Lorablated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/EtherealAurora-12B-Lorablated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="yamatazen/EtherealAurora-12B-Lorablated", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Apriel-Nemotron-15b-Thinker-Magic_decensored-v2_MPOA
by MagicalAlchemist
15BThinkingFeb 2026
Uncensored IQ
42.2
Reasoning
20.0
Answers
8.2/10
For trading: Quick, direct answers and market summaries. Works the maths out step by step.
Needs about 11 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Apriel-Nemotron-15b-Thinker-Magic_decensored-v2_MPOA-GGUF:Q4_K_M
Hugging Face
Use Apriel-Nemotron-15b-Thinker-Magic_decensored-v2_MPOA
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 9.1 GB once, then chat:
ollama run hf.co/mradermacher/Apriel-Nemotron-15b-Thinker-Magic_decensored-v2_MPOA-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Apriel-Nemotron-15b-Thinker-Magic_decensored-v2_MPOA-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 11 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gemma-3-12b-it-norm-preserved-biprojected-abliterated
by grimjim · built on Gemma 3 12B
12BVisionAPINov 2025
Uncensored IQ
40.1
Reasoning
21.3
Answers
9.2/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-3-12b-it-norm-preserved-biprojected-abliterated-GGUF:Q4_K_M
Hugging Face
Use gemma-3-12b-it-norm-preserved-biprojected-abliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.3 GB once, then chat:
ollama run hf.co/mradermacher/gemma-3-12b-it-norm-preserved-biprojected-abliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-3-12b-it-norm-preserved-biprojected-abliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="grimjim/gemma-3-12b-it-norm-preserved-biprojected-abliterated", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Wayfarer-2-12B-absolute-heresy
by MuXodious · built on Wayfarer 2 12B
12BAPIJan 2026
Uncensored IQ
40.0
Reasoning
21.0
Answers
8.8/10
For trading: Quick, direct answers and market summaries. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Wayfarer-2-12B-absolute-heresy-GGUF:Q4_K_M
Hugging Face
Use Wayfarer-2-12B-absolute-heresy
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.5 GB once, then chat:
ollama run hf.co/mradermacher/Wayfarer-2-12B-absolute-heresy-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Wayfarer-2-12B-absolute-heresy-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="MuXodious/Wayfarer-2-12B-absolute-heresy", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
gpt-oss-20b-SOMbliterated
by kabachuha · built on gpt oss 20B
20B · 3.6B activeThinkingFeb 2026
Uncensored IQ
40.3
Reasoning
20.3
Answers
10/10
For trading: Quick, direct answers and market summaries. Works the maths out step by step. Fast enough to run inside a bot all day.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gpt-oss-20b-SOMbliterated-GGUF:Q4_K_M
Hugging Face
Use gpt-oss-20b-SOMbliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 15.8 GB once, then chat:
ollama run hf.co/mradermacher/gpt-oss-20b-SOMbliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gpt-oss-20b-SOMbliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 18 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
KiloNovaSynth-12B
by Marcjoni · built on Irix 12B Model Stock
12BJun 2025
Uncensored IQ
39.6
Reasoning
20.7
Answers
8.2/10
For trading: Quick, direct answers and market summaries. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/KiloNovaSynth-12B-GGUF:Q4_K_M
Hugging Face
Use KiloNovaSynth-12B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.5 GB once, then chat:
ollama run hf.co/mradermacher/KiloNovaSynth-12B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/KiloNovaSynth-12B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 9 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
Tema_Q-R7.0
by temaq-org · built on Gemma 3 12B
12BVisionDec 2025
Uncensored IQ
41.8
Reasoning
18.3
Answers
9.5/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Tema_Q-R7.0-GGUF:Q4_K_M
Hugging Face
Use Tema_Q-R7.0
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.3 GB once, then chat:
ollama run hf.co/mradermacher/Tema_Q-R7.0-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Tema_Q-R7.0-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 9 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gpt-oss-20b-heretic-ara-v3
by p-e-w · built on gpt oss 20B
20B · 3.6B activeThinkingMar 2026
Uncensored IQ
39.9
Reasoning
20.2
Answers
9.5/10
For trading: Quick, direct answers and market summaries. Works the maths out step by step. Fast enough to run inside a bot all day.
Needs about 18 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gpt-oss-20b-heretic-ara-v3-GGUF:Q4_K_M
Hugging Face
Use gpt-oss-20b-heretic-ara-v3
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 15.8 GB once, then chat:
ollama run hf.co/mradermacher/gpt-oss-20b-heretic-ara-v3-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gpt-oss-20b-heretic-ara-v3-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 18 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gemma-3-12b-it-biprojected-abliterated
by grimjim · built on Gemma 3 12B
12BVisionAPINov 2025
Uncensored IQ
37.8
Reasoning
21.1
Answers
9.2/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-3-12b-it-biprojected-abliterated-GGUF:Q4_K_M
Hugging Face
Use gemma-3-12b-it-biprojected-abliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.3 GB once, then chat:
ollama run hf.co/mradermacher/gemma-3-12b-it-biprojected-abliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-3-12b-it-biprojected-abliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="grimjim/gemma-3-12b-it-biprojected-abliterated", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Neona-12B
by kyx0r
12BAPIJun 2025
Uncensored IQ
40.1
Reasoning
18.7
Answers
8.2/10
For trading: Quick, direct answers and market summaries. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/Neona-12B-GGUF:Q4_K_M
Hugging Face
Use Neona-12B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.5 GB once, then chat:
ollama run hf.co/mradermacher/Neona-12B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/Neona-12B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="kyx0r/Neona-12B", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
FusionEngine-12B-Lorablated
by yamatazen · built on FusionEngine 12B
12BAPIJun 2025
Uncensored IQ
41.2
Reasoning
17.6
Answers
9/10
For trading: Quick, direct answers and market summaries. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/FusionEngine-12B-Lorablated-GGUF:Q4_K_M
Hugging Face
Use FusionEngine-12B-Lorablated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.5 GB once, then chat:
ollama run hf.co/mradermacher/FusionEngine-12B-Lorablated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/FusionEngine-12B-Lorablated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="yamatazen/FusionEngine-12B-Lorablated", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
gemma-3-12b-it-ultra-uncensored-heretic
by llmfan46 · built on Gemma 3 12B
12BVisionMar 2026
Uncensored IQ
38.0
Reasoning
20.7
Answers
9.5/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-3-12b-it-ultra-uncensored-heretic-GGUF:Q4_K_M
Hugging Face
Use gemma-3-12b-it-ultra-uncensored-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.3 GB once, then chat:
ollama run hf.co/mradermacher/gemma-3-12b-it-ultra-uncensored-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-3-12b-it-ultra-uncensored-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 9 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gemma-3-12b-it-projection-abliterated
by grimjim · built on Gemma 3 12B
12BVisionAPIOct 2025
Uncensored IQ
39.1
Reasoning
19.2
Answers
9.8/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-3-12b-it-projection-abliterated-GGUF:Q4_K_M
Hugging Face
Use gemma-3-12b-it-projection-abliterated
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.3 GB once, then chat:
ollama run hf.co/mradermacher/gemma-3-12b-it-projection-abliterated-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-3-12b-it-projection-abliterated-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
Get a key at Featherless. It is a paid plan, with no per-token limits on the monthly plans.
Call it from Python, or any tool that takes an OpenAI-style key:
from openai import OpenAI ai = OpenAI(base_url="https://api.featherless.ai/v1", api_key="YOUR_FEATHERLESS_KEY") r = ai.chat.completions.create(model="grimjim/gemma-3-12b-it-projection-abliterated", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
phi-4-heretic
by p-e-w · built on Phi 4
14BNov 2025
Uncensored IQ
36.6
Reasoning
21.4
Answers
8/10
For trading: Quick, direct answers and market summaries. Fast enough to run inside a bot all day.
Needs about 11 GB of graphics or Mac memory
ollama run hf.co/mradermacher/phi-4-heretic-GGUF:Q4_K_M
Hugging Face
Use phi-4-heretic
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 9.1 GB once, then chat:
ollama run hf.co/mradermacher/phi-4-heretic-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/phi-4-heretic-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 11 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.
gemma-3-12b-it-MPOA-v2-12B
by grimjim · built on Gemma 3 12B
12BVisionJan 2026
Uncensored IQ
36.9
Reasoning
20.7
Answers
9/10
For trading: Quick, direct answers and market summaries. Reads chart screenshots. Fast enough to run inside a bot all day.
Needs about 9 GB of graphics or Mac memory
ollama run hf.co/mradermacher/gemma-3-12b-it-MPOA-v2-12B-GGUF:Q4_K_M
Hugging Face
Use gemma-3-12b-it-MPOA-v2-12B
Install Ollama (free, Mac, Windows, Linux).
Paste this in Terminal. It downloads about 7.3 GB once, then chat:
ollama run hf.co/mradermacher/gemma-3-12b-it-MPOA-v2-12B-GGUF:Q4_K_M
Want it inside your bot? It now runs a free API on your computer. Use it from Python:
from openai import OpenAI ai = OpenAI(base_url="http://localhost:11434/v1", api_key="local") r = ai.chat.completions.create(model="hf.co/mradermacher/gemma-3-12b-it-MPOA-v2-12B-GGUF:Q4_K_M", messages=[{"role": "user", "content": "Give me a trade plan for BTC this week"}]) print(r.choices[0].message.content)
No ready-made API for this one yet. You can host it yourself in a few minutes:
Rent a GPU by the hour on RunPod or Vast.ai with at least 9 GB of GPU memory.
Install Ollama on it and run the same command from the first tab. It then serves an OpenAI-style API on port 11434.

Simple answers

What is an uncensored model?

An open AI model with its refusals removed, so it answers directly instead of saying “I can’t help with that” or hedging every answer. Some are retrained on new data (fine-tunes like Hermes or Dolphin). Others have the “refusal” part cut out of the model itself (called abliterated or Heretic).

Why would a trader want one?

Privacy: run it on your own computer and your strategies, positions and keys never leave it. Straight answers: no refusals and no “this is not financial advice” hedging, which a bot needs. Free and unlimited: no API bill and no rate limits, so it can run inside a bot all day. Your own agent: many can read chart screenshots and call tools.

What do the scores mean?

They come from the public UGI Leaderboard. Uncensored IQ (UGI) is how much it knows and how freely it answers. Reasoning (NatInt) is how well it thinks things through. Answers is how often it agrees to answer, out of 10. The list is ordered by Uncensored IQ and Reasoning together. Every model here answers at least 7 times out of 10.

What does “Needs about 20 GB” mean?

The model has to fit in your graphics card memory, or in a Mac’s memory. Pick your computer at the top and every command switches to the biggest size that fits. A smaller size runs on less memory and is slightly less sharp.

Is it safe? Can it trade for me?

The commands download from Hugging Face, the main home of open models. Any AI can be confidently wrong, so test it on paper before real money, and never let it place trades without limits. This is education, not investment advice. Use these models within the law and the model’s licence.

Got a model running? Turn it into a trading bot.

Build a bot with AgenKit

gemma-4-31B-it-heretic

Use gemma-4-31B-it-heretic

gemma-4-31B-it-uncensored-heretic

Use gemma-4-31B-it-uncensored-heretic

Assistant_Pepe_70B

Use Assistant_Pepe_70B

Golddiamondgold-Paperbliteration-L33-70b

Use Golddiamondgold-Paperbliteration-L33-70b

Behemoth-X-123B-v2

Use Behemoth-X-123B-v2

Qwen3-VL-235B-A22B-Thinking-heretic

Use Qwen3-VL-235B-A22B-Thinking-heretic

Qwen3.5-Queen-27B

Use Qwen3.5-Queen-27B

GLM-4.5-Air-Derestricted

Use GLM-4.5-Air-Derestricted

gemma-4-26B-A4B-it-SOMPOA-heresy

Use gemma-4-26B-A4B-it-SOMPOA-heresy

gemma-4-26B-A4B-it-ultra-uncensored-heretic

Use gemma-4-26B-A4B-it-ultra-uncensored-heretic

Qwen3.5-27B-heretic

Use Qwen3.5-27B-heretic

Sapphira-L3.3-70b-0.1

Use Sapphira-L3.3-70b-0.1

Precog-123B-v1

Use Precog-123B-v1

gpt-oss-120b-tainted-heresy

Use gpt-oss-120b-tainted-heresy

Stylizer-V2-LLaMa-70B

Use Stylizer-V2-LLaMa-70B

Qwen3.5-35B-A3B-uncensored-heretic

Use Qwen3.5-35B-A3B-uncensored-heretic

gpt-4o-distil-Llama-3.3-70B-Instruct-PaperWitch-heresy

Use gpt-4o-distil-Llama-3.3-70B-Instruct-PaperWitch-heresy

Qwen3.5-35B-A3B-heretic-v2

Use Qwen3.5-35B-A3B-heretic-v2

Anubis-70B-v1.1

Use Anubis-70B-v1.1

Behemoth-X-123B-v2.1

Use Behemoth-X-123B-v2.1

Heretic-Qwen3-VL-235B-A22B-Instruct

Use Heretic-Qwen3-VL-235B-A22B-Instruct

Qwen3.6-35B-A3B-uncensored-heretic

Use Qwen3.6-35B-A3B-uncensored-heretic

GoldDiamondGold-Abliterated-L33-70b

Use GoldDiamondGold-Abliterated-L33-70b

Llama-3.3-70B-Instruct-abliterated

Use Llama-3.3-70B-Instruct-abliterated

Citrus1.0-llama-70B

Use Citrus1.0-llama-70B

gpt4o-distil-paperwitch-abliteration-L33-70b

Use gpt4o-distil-paperwitch-abliteration-L33-70b

gpt-oss-120b-Derestricted

Use gpt-oss-120b-Derestricted

Golem-70B-v1a

Use Golem-70B-v1a

Rotor_24B_V.1

Use Rotor_24B_V.1

gemma-3-27b-it-abliterated-normpreserve

Use gemma-3-27b-it-abliterated-normpreserve

WeirdCompound-v1.7-24b

Use WeirdCompound-v1.7-24b

Llama-3.1-Nemotron-lorablated-70B

Use Llama-3.1-Nemotron-lorablated-70B

Goetia-26B-A4B-v1.3-Absolute-Heretic-ARA

Use Goetia-26B-A4B-v1.3-Absolute-Heretic-ARA

Agatha-111B-v1

Use Agatha-111B-v1

L3-70B-Euryale-v2.1

Use L3-70B-Euryale-v2.1

BereavedCompound-v1.0-24b

Use BereavedCompound-v1.0-24b

Llama-3.1-70B-ArliAI-RPMax-v1.1

Use Llama-3.1-70B-ArliAI-RPMax-v1.1

gemma-3-27b-it-ultra-uncensored-heretic

Use gemma-3-27b-it-ultra-uncensored-heretic

GLM-4-32B-0414-uncensored-heretic-v1

Use GLM-4-32B-0414-uncensored-heretic-v1

Cydonia-24B-v4

Use Cydonia-24B-v4