Every run: the agent's full trajectory, its submissions over the 20-hour budget, and the final score.
| # | Model | Reward | Cost | In tok | Out tok | Time | Steps | Submits |
|---|---|---|---|---|---|---|---|---|
| 1 | Gemini 3.8 Flash | 0.4553 | $73.38 | 570.9M | 924k | 15h 34m | 1436 | 4 |
| 2 | Gemini 3.8 Flash | 0.4390 | $64.68 | 501.7M | 861k | 10h 34m | 1229 | 2 |
| 3 | Gemini 3.8 Flash | 0.3577 | $44.62 | 399.0M | 781k | 9h 40m | 979 | 7 |
| 4 | Gemini 3.8 Flash | 0.1301 | $76.16 | 654.7M | 1.0M | 15h 23m | 1694 | 6 |
| 5 | Gemini 3.8 Flash | 0.1057 | $39.41 | 364.3M | 690k | 7h 35m | 868 | 4 |