GPT-6.2 models top AA-Robotics v0.1 with scores of 79 and 78

ArtificialAnlys · x · 2026-11-18

Artificial Analysis' AA-Robotics v0.1 research preview, measuring how well language models understand, plan and act in the real world, is led by GPT-6.2 Astra (79) and GPT-6.2 Sol (78), with Claude Fable 5.2 at 75. Gemini 4 Hydrogen (58) and Argon (55) are the best non-OpenAI/Anthropic entries; Claude Opus 5.5 manages only 49.

Original post →

More from Embodied

Embodied channel →