704-vote robotics Elo ranks Claude Opus 5.5 first at 1379; no model completes the task
qinzytech · x · 2026-10-21
Embodied Analysis published overall Elo scores from 704 votes in its robotics evaluation: Claude Opus 5.5 at 1379, Gemini 3.8 Flash 1218.4, GPT-6 Sol 1121.3, GPT-6 Astra 1099.5.
All models attempted the same simulated task (place a croissant on the cutting board, then bring a jar of jam):
- Opus 5.5: places the croissant, retrieves the jam, but never places the jar on the board.
- Gemini 3.8 Flash: places the croissant (4s playback), moves toward the jam; jar not placed.
- GPT-6 Sol: repeatedly adjusts around the croissant without placing it.
- GPT-6 Astra: drops the croissant on the floor and approaches the cabinet without recovering it.
No single run completed both steps, underscoring how far frontier models remain from reliable embodied manipulation.
More from Embodied
- Florida debuts America's first manned electric flying police vehicle for surveillance — Polymarket · 2026-10-21
- ETH's HumanoidPose recovers full-body robot pose from plain video, no internal sensors — ETHZurich · 2026-10-21
- Chris Paxton: the whole robotics stack won't run in the cloud — chris_j_paxton · 2026-10-21
- Robotics roundtable: data flywheels for real-world robot autonomy — chris_j_paxton · 2026-10-21
- Humanoid robot hands report: durability, not dexterity, gates the market at 200-1,250 hrs vs 10,000-hr target — MarwaEldiwiny · 2026-10-21
- Recording calls with PLAUD Notepin and piping transcripts into AI threads — curious_vii · 2026-10-21