WebsiteBenchWebsiteBench

TaskRabbit

Rebuilt by Claude Opus 5.5: 237 of 871 hidden tests passed (27.2%)

reference.local/
Reference for TaskRabbit, HomeReference
rebuild.local/
Rebuild by Claude Opus 5.5, HomeClaude Opus 5.5
    Every model on TaskRabbit
    ModelPassed%Screenshot
    Claude Opus 5.523727.2
    Qwen 3.8 Max19422.3
    GLM 5.3 Flash16819.3
    GPT 6.1 Sol16819.3
    Claude Opus 515517.8
    DeepSeek V4.1 Flash15517.8
    Grok 4.613815.8
    Kimi K3Timeout

    Every result in the WaffleRun the benchmark