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 2026-11-12
#ModelRewardCostIn tokOut tokTimeStepsSubmits
1Gemini 4 Carbon0.3810$31.5327.6M200k19h 50m1275
2Gemini 4 Carbon0.3810$39.0042.5M193k19h 50m1784
3Gemini 4 Carbon0.3333$35.1531.4M180k19h 44m1365
4Gemini 4 Carbon0.1429$44.5444.2M256k19h 38m1794
5Gemini 4 Carbon0.1190$37.4240.0M214k19h 50m1655