GPT-6 Astra appears to show a "step change" in spatial reasoning based on early benchmarks

GPT-6 Astra has achieved significant advancements in spatial understanding according to new robotics benchmarks. On the StationeryBench platform, the model completed 7 out of 100 tasks using dual-arm robots, while competitor MolmoAct2 failed to finish any tasks. This achievement is considered a 'step change' in spatial reasoning. The results indicate that GPT-6 Astra significantly outperforms its competitors in spatial understanding tasks, reflecting substantial improvements in the model's capabilities. Researchers suggest that this progress could enhance practical applications across various fields, including intelligent robotics.