EmbodiedSWE wins CoRL 2026 Best Paper: coding agents generate robot training data
PeterHndrsn · x · 2026-11-14
EmbodiedSWE won Best Paper at the CoRL 2026 Workshop on Agentic Robotics, studying how frontier coding agents can help robotics. Four components:
- EmbodiedSWE-Bench: an agent-native benchmark of long-horizon dexterous everyday tasks built on Isaac Lab
- Evaluation of frontier coding agents on task performance, completion time, and inference cost
- EmbodiedSWE-Gen: diversifies one verified agent solution into a large trajectory dataset for training general robot policies
- Agent improvement via generating new tasks from existing ones and RL on verified outcomes
The work realizes the Real2Sim2Real pipeline through general coding agents that RL researchers envisioned years ago. Code and paper are open source (193 GitHub stars).
Related event: EmbodiedSWE wins Best Paper at CoRL 2026 Agentic Robotics Workshop(2 posts)→
More from Embodied
- Figure AI seen most exposed to Nvidia fallout: $3.5-6B compute deal in limbo — kscottz · 2026-11-14
- 5,400 real-robot rollouts benchmark Claude Fable 5.2 at 65%, ahead of Gemini and GPT — ricklamers · 2026-11-14
- Ducted-Fan Robot Uses Coandă Effect for 3D Thrust Vectoring and Stable Flight — bennash · 2026-11-14
- ASUS teases its first humanoid robot in near-black teaser video — Distinct-Question-16 · 2026-11-13
- Kodiak's in-house sim stack BreakPoint cuts costs to 2% while running 16x faster — Scobleizer · 2026-11-13
- YC startup Antimatter builds voice-first hardware after AirPods misheard 30% of words — aakashgupta · 2026-11-13