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 | 1.0000 | $164.97 | 80.6M | 296k | 2h 47m | 306 | 8 |
| 2 | GPT-6 Astra | 1.0000 | $50.75 | 31.1M | 211k | 1h 50m | 184 | 6 |
| 3 | GPT-6 Astra | 0.9768 | $52.93 | 33.2M | 214k | 1h 55m | 200 | 5 |
| 4 | GPT-6 Astra | 0.4310 | $132.03 | 68.3M | 275k | 2h 40m | 289 | 5 |
| 5 | GPT-6 Astra | 0.2020 | $84.84 | 43.9M | 239k | 1h 53m | 208 | 6 |