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 | $46.19 | 26.4M | 213k | 1h 23m | 134 | 7 |
| 2 | GPT-6 Astra | 1.0000 | $35.97 | 23.6M | 186k | 1h 21m | 133 | 4 |
| 3 | GPT-6 Astra | 1.0000 | $131.76 | 64.1M | 279k | 2h 36m | 232 | 7 |
| 4 | GPT-6 Astra | 0.9976 | $79.59 | 42.6M | 240k | 2h 8m | 195 | 9 |
| 5 | GPT-6 Astra | 0.9976 | $93.69 | 48.6M | 237k | 1h 56m | 206 | 6 |