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 Fable 5 | 0.8760 | $652.52 | 521.9M | 1.8M | 18h 59m | 1216 | 15 |
| 2 | Claude Fable 5 | 0.8612 | $916.45 | 747.8M | 2.3M | 18h 51m | 1719 | 16 |
| 3 | Claude Fable 5 | 0.7325 | $725.08 | 587.9M | 1.9M | 18h 21m | 1299 | 12 |
| 4 | Claude Fable 5 | 0.7151 | $484.68 | 385.7M | 1.3M | 18h 44m | 837 | 11 |
| 5 | Claude Fable 5 | 0.6534 | $728.09 | 583.8M | 2.0M | 17h 53m | 1264 | 6 |