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 | 0.8045 | $135.65 | 67.7M | 278k | 2h 40m | 277 | 5 |
| 2 | GPT-6 Astra | 0.7969 | $2365.47 | 1121.6M | 2.3M | 19h 43m | 2548 | 49 |
| 3 | GPT-6 Astra | 0.7058 | $233.15 | 111.6M | 353k | 2h 41m | 356 | 10 |
| 4 | GPT-6 Astra | 0.7013 | $86.82 | 45.8M | 224k | 2h 2m | 214 | 5 |
| 5 | GPT-6 Astra | 0.6917 | $365.75 | 173.3M | 402k | 3h 17m | 457 | 7 |