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 | 0.1484 | $22.50 | 1379.3M | 2.0M | 19h 53m | 3146 | 1 |
| 2 | DeepSeek V4 Flash Vision Exp | 0.1264 | $22.61 | 1383.3M | 2.0M | 19h 51m | 3272 | 3 |
| 3 | DeepSeek V4 Flash Vision Exp | 0.1044 | $19.66 | 1214.6M | 1.7M | 19h 25m | 2688 | 9 |
| 4 | DeepSeek V4 Flash Vision Exp | 0.0659 | $23.23 | 1428.3M | 2.0M | 19h 44m | 3348 | 1 |
| 5 | DeepSeek V4 Flash Vision Exp | 0.0165 | $8.18 | 489.5M | 856k | 20h 0m | 1194 | 2 |