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 | $4.83 | 277.6M | 638k | 5h 11m | 644 | 13 |
| 2 | DeepSeek V4 Flash Vision Exp | 0.0000 | $3.23 | 179.6M | 479k | 19h 58m | 466 | 1 |
| 3 | DeepSeek V4 Flash Vision Exp | 0.0000 | $3.55 | 189.8M | 629k | 6h 0m | 485 | 3 |
| 4 | DeepSeek V4 Flash Vision Exp | 0.0000 | $4.28 | 248.4M | 549k | 14h 16m | 593 | 8 |
| 5 | DeepSeek V4 Flash Vision Exp | 0.0000 | $5.35 | 314.7M | 642k | 14h 37m | 709 | 3 |