Skip to content

Place dual shoes unlimited

Codex + GPT-6 Luna, reasoning xhigh (Azure OpenAI API, 2026-09-30 stress test)

← Task page · All RoboTwin 2.0 tasks in this run

ⓘ Note: regraded offline: the verifier ran out of GPU memory (cuRobo) while other jobs shared one of our machines machine's GPU, so it wrote grader_error; the same grader rerun on the collected trajectory, alone on the GPU, reads success (deterministic).

Model use: est. $0.062 at list price