GPT-6 Astra and Claude Opus 5 nearly tied on embodied task: 34-6 over 40 votes, 16-point Elo gap
qinzytech · x · 2026-10-13
Embodied Analysis reports GPT-6 Astra and Claude Opus 5 are nearly tied on its embodied benchmark: both went 34-6 across 40 votes, with Astra holding a 16-point Elo lead (1222 vs 1206).
The task required every object to be placed in a box with its front side facing left. In a 2.6x-speed video, Opus 5 cleared the table and returned both arms to rest by 0:19, while Astra also cleared the table but kept one arm moving in and around the box until the clip ends at 0:20.
The reviewer notes the objects are visible in the box but their final orientation is hard to verify, and that a full assessment should weigh placement, orientation, and stopping behavior separately.
More from Embodied
- RoboARCade launches: agents directly control robots, best GPT-6 setup tops out at 49.8% success — leto__jean · 2026-10-13
- Free-Init: scan-free, motion-free initialization for Doppler LiDAR-inertial systems — rsasaki0109 · 2026-10-13
- On-device robot practices soccer throw-ins with no external tracking — chris_j_paxton · 2026-10-13
- Custom ComfyUI node repurposes Mixamo's 3D rigs to pose 2D characters — Willonfire8 · 2026-10-13
- First Origami Challenge at IROS 2026: 14 teams push dexterous manipulation to the limit — chris_j_paxton · 2026-10-13
- Robotic hand learns to crawl by treating its fingers as legs — LukeW · 2026-10-13