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 | 0.0891 | $17.04 | 1034.3M | 1.6M | 6h 42m | 2273 | 6 |
| 2 | DeepSeek V4 Flash Vision Exp | 0.0545 | $16.46 | 992.0M | 1.7M | 6h 43m | 2222 | 11 |
| 3 | DeepSeek V4 Flash Vision Exp | 0.0000 | $13.51 | 827.8M | 1.3M | 5h 43m | 1720 | 13 |
| 4 | DeepSeek V4 Flash Vision Exp | 0.0000 | $21.93 | 1323.6M | 2.2M | 8h 50m | 3095 | 7 |
| 5 | DeepSeek V4 Flash Vision Exp | 0.0000 | $33.36 | 1965.0M | 3.8M | 19h 31m | 4102 | 48 |