Skip to content

Stationerybench

Topic archive • 1 matches

Back to home • GEO summary endpoint

2026-09-13

Technology

  • GPT-6 Astra completes 7 of 100 robotics tasks in spatial benchmark: GPT-6 Astra showed significant gains in spatial understanding on the new StationeryBench robotics benchmark. The model successfully completed 7 out of 100 tasks using dual-arm robots, whereas the competitor model MolmoAct2 failed to complete any tasks.

    AI Models • The Decoder

    Permalink