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
1Muse Spark 1.20.3810$64.94126.7M662k16h 40m38627
2Muse Spark 1.20.3095$11.9541.8M358k12h 44m19924
3Muse Spark 1.20.2619$355.312070.9M1.9M15h 46m592022
4Muse Spark 1.20.1905$5.5220.2M214k2h 15m1353
5Muse Spark 1.20.0476$35.64132.6M581k10h 53m45113