GPT-6-Astra aces simulation benchmarks but struggles on real robots

A joint evaluation across 12 simulation benchmarks and 8 real robot platforms found GPT-6-Astra approaching 100% success in simulation but failing on real hardware, exposing a large sim-to-real gap in embodied AI scores.

2026-10-27 ~ 2026-10-29 · 2 related posts

• GPT-6-Astra embodied intelligence eval: near-100% on several of 12 benchmarks — chris_j_paxton · 2026-10-27
• 100% in sim, 0/11 real robot inserts: joint team exposes benchmark inflation in GPT-6-Astra eval — jiqizhixin · 2026-10-29