Traces
Every run: the agent's full trajectory, its submissions over the 20-hour budget, and the final score.
FFmpeg libswscale Optimization
task page →5 runs · indexed 2026-09-02| # | Model | Reward | Cost | In tok | Out tok | Time | Steps | Submits |
|---|
| 1 | GPT-5.6 | 0.0000 | $66.82 | 75.1M | 179k | 1h 59m | 318 | 8 |
| 2 | GPT-5.6 | 0.0000 | $78.43 | 84.8M | 190k | 2h 4m | 348 | 5 |
| 3 | GPT-5.6 | 0.0000 | $123.31 | 124.9M | 233k | 2h 12m | 422 | 3 |
| 4 | GPT-5.6 | 0.0000 | $76.67 | 86.4M | 186k | 3h 5m | 369 | 10 |
| 5 | GPT-5.6 | 0.0000 | $117.96 | 120.8M | 233k | 3h 42m | 405 | 8 |