Traces

Every run: the agent's full trajectory, its submissions over the 20-hour budget, and the final score.

Multi-GPU Efficient Finetuning

task page →5 runs · indexed 2026-09-02
#ModelRewardCostIn tokOut tokTimeStepsSubmits
1GPT-5.60.2381$19.3323.5M110k15h 47m1673
2GPT-5.60.0238$19.6525.0M71k12h 53m1333
3GPT-5.60.0000$16.4525.9M78k16h 50m1802
4GPT-5.60.0000$24.1334.7M95k16h 10m1812
5GPT-5.60.0000$30.4139.1M90k14h 8m1993