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 | 0.4570 | $770.42 | 373.2M | 731k | 19h 37m | 1013 | 12 |
| 2 | GPT-6 Astra | 0.3866 | $1781.91 | 865.7M | 1.5M | 19h 55m | 2133 | 25 |
| 3 | GPT-6 Astra | 0.2803 | $1564.06 | 755.2M | 1.2M | 19h 40m | 1857 | 16 |
| 4 | GPT-6 Astra | 0.1048 | $823.35 | 394.4M | 766k | 19h 59m | 1018 | 14 |
| 5 | GPT-6 Astra | 0.0879 | $1737.29 | 826.0M | 1.5M | 19h 53m | 1937 | 16 |