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 | DeepSeek V4 Flash Vision Exp | 0.4014 | $18.12 | 1122.6M | 1.6M | 19h 51m | 2602 | 23 |
| 2 | DeepSeek V4 Flash Vision Exp | 0.2342 | $7.03 | 436.2M | 602k | 8h 13m | 914 | 3 |
| 3 | DeepSeek V4 Flash Vision Exp | 0.2234 | $21.23 | 1304.8M | 2.0M | 15h 38m | 2836 | 7 |
| 4 | DeepSeek V4 Flash Vision Exp | 0.1638 | $15.31 | 956.5M | 1.3M | 19h 4m | 2023 | 3 |
| 5 | DeepSeek V4 Flash Vision Exp | 0.0766 | $15.69 | 953.5M | 1.5M | 16h 48m | 2199 | 5 |