⬢github · 334 ★ +105 since we first saw it · pushed 23 h ago
JoasASantos/Offensive-Security-AI-Models
Uncensored AI models or those fine-tuned for cybersecurity tasks.
A curated, benchmark-style list of open-weight LLMs that are uncensored or fine-tuned for offensive and defensive security work — penetration testing, red teaming, vuln research, and SOC tasks. Each entry includes specs like base model, parameter count, context length, VRAM needs, training data, uncensoring method, and download links from HuggingFace.
Why now: It was featured in a Hacker News discussion ('Uncensored and Offensive Security AI Models Benchmark') and recently updated, offering a side-by-side comparison of these models' capabilities and hardware requirements.
Who it is for: Security researchers, red teamers, and pentesters with authorization who want to run or compare specialized open-weight models locally.
Stars over our 21 snapshots: 229 to 334, since 5 h ago.
Where people talked about it
API: https://socialmediatrends-api.osmike.com/v1/repos/JoasASantos/Offensive-Security-AI-Models