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 | 0.8707 | $2626.81 | 1259.5M | 2.3M | 18h 25m | 2839 | 58 |
| 2 | GPT-6 Astra | 0.8658 | $272.49 | 128.6M | 369k | 2h 40m | 367 | 7 |
| 3 | GPT-6 Astra | 0.8334 | $251.26 | 124.6M | 331k | 2h 35m | 409 | 4 |
| 4 | GPT-6 Astra | 0.8266 | $344.95 | 163.8M | 406k | 3h 8m | 459 | 8 |
| 5 | GPT-6 Astra | 0.8198 | $254.82 | 122.3M | 354k | 2h 28m | 374 | 7 |