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 | GPT-5.6 Sol | 0.7649 | $74.74 | 74.2M | 207k | 2h 36m | 270 | 5 |
| 2 | GPT-5.6 Sol | 0.6303 | $111.93 | 116.6M | 261k | 11h 2m | 425 | 3 |
| 3 | GPT-5.6 Sol | 0.6093 | $23.53 | 32.7M | 162k | 1h 28m | 207 | 3 |
| 4 | GPT-5.6 Sol | 0.4658 | $119.15 | 122.5M | 308k | 3h 17m | 463 | 6 |
| 5 | GPT-5.6 Sol | 0.3383 | $180.46 | 175.3M | 266k | 4h 7m | 474 | 10 |