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 | 0.8708 | $2.68 | 149.8M | 376k | 16h 23m | 473 | 5 |
| 2 | DeepSeek V4 Flash Vision Exp | 0.7876 | $3.53 | 195.8M | 540k | 19h 32m | 519 | 8 |
| 3 | DeepSeek V4 Flash Vision Exp | 0.7578 | $8.78 | 501.2M | 1.2M | 13h 5m | 1157 | 3 |
| 4 | DeepSeek V4 Flash Vision Exp | 0.7467 | $8.27 | 472.9M | 1.1M | 11h 58m | 1140 | 23 |
| 5 | DeepSeek V4 Flash Vision Exp | 0.4680 | $3.00 | 159.1M | 520k | 6h 30m | 442 | 5 |