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 | 0.0668 | $8.52 | 493.3M | 1.0M | 6h 44m | 1095 | 8 |
| 2 | DeepSeek V4 Flash Vision Exp | 0.0636 | $9.11 | 528.7M | 1.1M | 6h 10m | 1172 | 5 |
| 3 | DeepSeek V4 Flash Vision Exp | 0.0602 | $13.77 | 800.3M | 1.6M | 9h 14m | 1831 | 19 |
| 4 | DeepSeek V4 Flash Vision Exp | 0.0483 | $9.39 | 554.1M | 1.0M | 5h 54m | 1197 | 16 |
| 5 | DeepSeek V4 Flash Vision Exp | 0.0000 | $12.34 | 718.0M | 1.4M | 8h 32m | 1678 | 16 |