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 | 0.7901 | $2673.96 | 1271.2M | 2.5M | 19h 58m | 2991 | 31 |
| 2 | GPT-6 Astra | 0.7705 | $2639.52 | 1259.7M | 2.5M | 19h 50m | 2961 | 27 |
| 3 | GPT-6 Astra | 0.7148 | $2733.41 | 1304.4M | 2.5M | 19h 53m | 2977 | 29 |
| 4 | GPT-6 Astra | 0.6873 | $2846.47 | 1353.6M | 2.5M | 19h 54m | 3114 | 36 |
| 5 | GPT-6 Astra | 0.6437 | $2584.51 | 1240.4M | 2.4M | 19h 47m | 2920 | 23 |