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-6 Astra | 0.9350 | $571.91 | 270.6M | 494k | 3h 53m | 570 | 4 |
| 2 | GPT-6 Astra | 0.9350 | $540.43 | 256.4M | 446k | 3h 37m | 544 | 4 |
| 3 | GPT-6 Astra | 0.8699 | $378.55 | 179.5M | 449k | 3h 11m | 459 | 5 |
| 4 | GPT-6 Astra | 0.8699 | $279.05 | 132.3M | 363k | 2h 37m | 372 | 4 |
| 5 | GPT-6 Astra | 0.8293 | $299.55 | 142.9M | 339k | 2h 28m | 402 | 4 |