| name: WildClawBench | |
| description: Hard, practical, end-to-end evaluation for AI agents — in the wild. | |
| evaluation_framework: wildclawbench | |
| tasks: | |
| - id: overall | |
| name: Overall | |
| config: default | |
| split: test | |
| - id: avg_time | |
| name: Avg Time (min) | |
| config: default | |
| split: test | |
| - id: avg_cost | |
| name: Avg Cost ($) | |
| config: default | |
| split: test |