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

#	Model	Reward	Cost	In tok	Out tok	Time	Steps	Submits
1	GLM-5.3	0.2857	$12.88	26.8M	144k	19h 0m	234	2
2	GLM-5.3	0.2857	$10.42	18.4M	127k	19h 10m	175	6
3	GLM-5.3	0.0714	$17.61	33.2M	221k	18h 56m	211	2
4	GLM-5.3	0.0000	$20.26	34.9M	208k	19h 31m	211	4
5	GLM-5.3	0.0000	$17.67	32.5M	199k	19h 29m	215	2