Pirate Face Rescues 1,240 LLM Models from Deletion

Pirate Face has successfully rescued 1,240 large language models (LLMs) from potential deletion, utilizing a decentralized infrastructure aimed at preserving open-source AI. The platform allows users to browse, download, and seed models without requiring an account, promoting accessibility.

Notable models updated recently include:

• prism-ml/Ternary-Bonsai-2-27B-gguf (54 GB) updated 4 days ago.
• sentence-transformers/all-MiniLM-L6-v2 (91 MB) updated today.
• convaiinnovations/laya (843 MB) updated 2 days ago.
• XingChen-AGI/Xing4.0-29B-A4B (62 GB) updated 3 days ago.
• deepseek-ai/DeepSeek-V4.1-Flash (765 GB) updated 11 days ago.
• Qwen/Qwen3.8-27B (56 GB) updated 38 days ago.
• baidu/Unlimited-OCR (6.7 GB) updated today.

Each model is verified against Hugging Face's official SHA-256 checksum, ensuring integrity and authenticity. Pirate Face claims that every model acts as a torrent that also downloads directly from Hugging Face, reinforcing their commitment to permanence for open-source AI.

Users earn 25 points for each distinct model rescued and 50 points for creating a new account. This initiative follows previous efforts to enhance performance and efficiency in AI data centers, as noted in Turing Wire's recent coverage. The ability to rescue and verify models could significantly impact how practitioners manage and utilize open-source AI resources.