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 | 0.9864 | $3111.97 | 1407.9M | 3.4M | 19h 58m | 3119 | 48 |
| 2 | GPT-6 Astra | 0.9049 | $3006.21 | 1380.7M | 3.3M | 20h 0m | 3129 | 60 |
| 3 | GPT-6 Astra | 0.8950 | $2950.79 | 1350.4M | 3.3M | 19h 50m | 3031 | 52 |
| 4 | GPT-6 Astra | 0.8938 | $3057.89 | 1408.3M | 3.3M | 19h 59m | 3204 | 96 |
| 5 | GPT-6 Astra | 0.8893 | $3214.36 | 1493.6M | 3.2M | 19h 59m | 3347 | 99 |