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 | 0.9807 | $122.08 | 64.3M | 290k | 2h 50m | 282 | 4 |
| 2 | GPT-6 Astra | 0.9807 | $62.98 | 37.7M | 245k | 1h 38m | 203 | 5 |
| 3 | GPT-6 Astra | 0.9807 | $123.05 | 61.8M | 300k | 2h 8m | 260 | 3 |
| 4 | GPT-6 Astra | 0.9791 | $34.15 | 22.2M | 179k | 1h 14m | 133 | 3 |
| 5 | GPT-6 Astra | 0.9791 | $174.62 | 85.4M | 285k | 2h 50m | 311 | 4 |