WildAWM synthesizes computer-use agent environments with open models
It yields 341 tasks that can exceed 100 steps; GPT-6-Astra solves 60%, and Qwen models trained only there improve on eight benchmarks.
2 stories, newest first · one card per story
It yields 341 tasks that can exceed 100 steps; GPT-6-Astra solves 60%, and Qwen models trained only there improve on eight benchmarks.
Hugging Face publishes observations on the state of the open-weights model ecosystem as of summer 2026.
Today 251 readers · 270 views · 20 reads · all time 108K views