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 | 0.9884 | $148.01 | 72.8M | 245k | 17h 13m | 283 | 8 |
| 2 | GPT-6 Astra | 0.9884 | $97.13 | 52.7M | 210k | 16h 8m | 257 | 6 |
| 3 | GPT-6 Astra | 0.9842 | $127.33 | 65.0M | 249k | 17h 10m | 287 | 7 |
| 4 | GPT-6 Astra | 0.9702 | $43.87 | 26.5M | 154k | 17h 20m | 152 | 9 |
| 5 | GPT-6 Astra | 0.9688 | $90.01 | 51.8M | 262k | 15h 58m | 286 | 7 |