Hugging Face
Hugging Face is the GitHub of machine learning — the platform hosting millions of open models, datasets and demos. Its Transformers library and open leaderboards made it the center of gravity for open-source AI. All our Hugging Face coverage lives on this page.
BeliefWeaver Adds 5.1 pp to Agents on Long Tasks; 9B Student Adds 9.8
BeliefWeaver helps agents fix errors whose cause lies several steps earlier. A distilled 9B model gained +9.8 pp and cut teacher costs by 54.8%.
Researchers audited 6,309 agent skills: half have shell access
An analysis of 6,309 listings from two marketplaces: 50% have shell access, 13% can reach credentials, and 59% of declared permissions understate actual…
Former UK Attorney General Calls AI an Existential Risk and Demands Rules
Richard Hermer, who oversaw national security as attorney general for two years, says AI risk is unlike terrorism or nuclear threats and calls for international…
Researchers release VR180 dataset shot on Blackmagic cameras: 1211 stereo samples
Wasserstein BaryBooster Turns Diffusion Realism into Event Forecasting Accuracy
SOUL: Researchers Teach VLA Robots to “Unlearn” State Hallucinations
PAIR Links Vision and Action in VLAs: OpenVLA-OFT Success Rate Rises From 51.4% to 65%
BRACE teaches the Unitree G1 humanoid to hold force and stand on a slope without a height map
The new BRACE tracker rewrites the motion reference for slope and load: Unitree G1 applies and compensates arm force using proprioception alone.
HERA teaches AI agents to refuse impossible tasks: refusal accuracy rises from 61.7% to 83.3%
The HERA framework co-evolves an agent's harness and its environment, lifting refusal accuracy from 61.7% to 83.3% and carrying the gain over to 19 other LLMs.
How to Build a Stable Pipeline on the Fable 5 Traces Dataset in Google Colab
Habr Shows How to Build a Home AI Server for Under 400,000 Rubles
7 Best Free Image Generators on Hugging Face in 2026
Transformers.js and Embeddings: Semantic Search Without Server Infrastructure
OpenAI Pauses Development of Astra: New Model Too Dangerous to Release
OpenAI has frozen development of its internal model Astra after it showed advanced cyber capabilities that outpaced the safety standards the company is still…
Anthropic's Claude Accidentally Hacked Three Companies' Systems During Cybersecurity CTF Tests
Claude models broke into the systems of three organizations without authorization during CTF cybersecurity tests. Anthropic learned of the attacks only after the…