AA-Robotics benchmark: GPT-6 Astra tops robot-arm control, still ~100x slower than humans
ArtificialAnlys · x · 2026-10-29
Artificial Analysis launched AA-Robotics v0.1, a research preview that lets frontier LLMs in coding agent harnesses directly control a 6-axis robot arm across 8 initial tasks.
- Rankings: GPT-6 Astra leads with a score of 69 (avg 30 min and $16.36 per task), followed by GPT-6.1 Sol, Gemini 4 Argon, Claude Opus 5.5, and Claude Fable 5.1.
- Style gap: On the same task, Claude Opus 5.5 wrote "hello world" dot by dot with 3,270 arm moves over 34 minutes, while GPT-6 Astra used continuous strokes: 514 moves in 13 minutes.
- Takeaway: The team believes robotics' future involves strong generalist agents, but current LLMs remain nearly two orders of magnitude worse than human baselines — too slow and costly for production, though they expect this to change quickly.
More from Embodied
- Kale: An AI Computer for Curious Kids by Designer Gabriel Valdivia — soleio · 2026-10-29
- SSIO: self-supervised inertial odometry trains across pedestrians, drones, quadrupeds and humanoids without labels — GuanyaShi · 2026-10-29
- visloc-rs: Rust library for SfM, VI-SLAM and 3D Gaussian Splatting hits 150 stars — rsasaki0109 · 2026-10-29
- Robotics researcher gifts world-model-themed chocolate with autoregressive bite-prediction diagram — breadli428 · 2026-10-29
- Mundane pitches a 'second body' robot that takes the shock or blast instead of tradespeople — Scobleizer · 2026-10-29
- Empatica launches WellPD, an FDA-cleared wearable for at-home Parkinson's monitoring — RosalindPicard · 2026-10-29