RoboColosseum launches open robot evals: 100+ robots, 10,000 free eval hours
DJiafei · x · 2026-11-06
Hugging Face, BitRobot Network and others launched RoboColosseum, an open robot eval initiative offering 100+ real robots (Unitree G1s, LeRobot SO-101s, YAMs, Franka and more), 140+ in-the-wild scenes, open rollout data, and 10,000 free eval hours.
The repost raises an intriguing question: can agentic-harness-style evaluation transfer to physical AI? Could a smaller model with the right tools beat Astra/Opus 5.5 on robot tasks — extending the LLM harness debate to robotics.
Related event: Hugging Face and Partners Launch RoboColosseum Open Real-Robot Evaluation(77 posts)→
More from Embodied
- RoboColosseum tests 5 VLA policies across 4 embodiments; environment diversity is the new bottleneck — DJiafei · 2026-11-06
- RoboColosseum: 20,000+ runs, 100+ robots, rankings flip by task — DJiafei · 2026-11-06
- Using Dance Choreography Tool Swing Choreo to Design Robot Motions — andrew_n_carr · 2026-11-06
- Cartwheel lands on founder-generated Emergent startup list for kinematic world models — andrew_n_carr · 2026-11-06
- Ai2 releases Molmo2-4B-tb, a vision-language planner for robot block stacking — DJiafei · 2026-11-06
- Fly connectome with 138k neurons drives a robot duck's escape reflexes in sim — willknight · 2026-11-06