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 | Claude Opus 5 | 0.8972 | $636.78 | 1144.6M | 1.6M | 19h 47m | 2383 | 4 |
| 2 | Claude Opus 5 | 0.8643 | $420.46 | 752.1M | 1.2M | 19h 22m | 1536 | 5 |
| 3 | Claude Opus 5 | 0.8376 | $279.11 | 485.1M | 948k | 19h 29m | 1196 | 3 |
| 4 | Claude Opus 5 | 0.8261 | $296.13 | 517.9M | 961k | 18h 59m | 1277 | 26 |
| 5 | Claude Opus 5 | 0.8114 | $315.87 | 552.3M | 1.0M | 19h 44m | 1262 | 18 |