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 | $7.62 | 422.1M | 1.1M | 12h 5m | 912 | 6 |
| 2 | DeepSeek V4 Flash Vision Exp | 0.0000 | $6.14 | 336.4M | 959k | 16h 29m | 853 | 10 |
| 3 | DeepSeek V4 Flash Vision Exp | 0.0000 | $4.17 | 227.6M | 660k | 6h 35m | 447 | 3 |
| 4 | DeepSeek V4 Flash Vision Exp | 0.0000 | $6.64 | 359.4M | 1.1M | 12h 22m | 882 | 5 |
| 5 | DeepSeek V4 Flash Vision Exp | 0.0000 | $6.97 | 389.0M | 1.0M | 14h 29m | 887 | 5 |