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 | 0.5582 | $23.11 | 1453.0M | 1.8M | 19h 40m | 3118 | 3 |
| 2 | DeepSeek V4 Flash Vision Exp | 0.4212 | $16.08 | 998.7M | 1.4M | 20h 0m | 2128 | 4 |
| 3 | DeepSeek V4 Flash Vision Exp | 0.3784 | $22.06 | 1380.9M | 1.8M | 19h 50m | 2999 | 12 |
| 4 | DeepSeek V4 Flash Vision Exp | 0.3253 | $27.96 | 1735.1M | 2.4M | 17h 46m | 3798 | 22 |
| 5 | DeepSeek V4 Flash Vision Exp | 0.3065 | $23.64 | 1484.7M | 1.9M | 18h 25m | 3157 | 30 |