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 | 0.9857 | $244.69 | 120.7M | 324k | 13h 51m | 408 | 10 |
| 2 | GPT-6 Astra | 0.9429 | $571.36 | 274.8M | 479k | 10h 52m | 637 | 7 |
| 3 | GPT-6 Astra | 0.8429 | $330.64 | 163.8M | 407k | 12h 13m | 526 | 9 |
| 4 | GPT-6 Astra | 0.8357 | $771.41 | 367.5M | 716k | 14h 57m | 903 | 12 |
| 5 | GPT-6 Astra | 0.7536 | $12.03 | 4.8M | 92k | 34m | 69 | 3 |