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 | 0.0000 | $11.65 | 704.6M | 1.2M | 10h 40m | 1549 | 15 |
| 2 | DeepSeek V4 Flash Vision Exp | 0.0000 | $19.65 | 1192.1M | 2.0M | 18h 3m | 2398 | 10 |
| 3 | DeepSeek V4 Flash Vision Exp | 0.0000 | $20.48 | 1199.6M | 2.4M | 11h 32m | 2563 | 9 |
| 4 | DeepSeek V4 Flash Vision Exp | 0.0000 | $21.51 | 1287.0M | 2.2M | 19h 10m | 2904 | 13 |
| 5 | DeepSeek V4 Flash Vision Exp | 0.0000 | $19.68 | 1182.5M | 2.0M | 19h 8m | 2617 | 23 |