WebsiteBenchWebsiteBench

TripIt

Rebuilt by GPT 6.1 Sol: 72 of 478 hidden tests passed (15.1%)

reference.local/
Reference for TripIt, HomeReference
rebuild.local/
Rebuild by GPT 6.1 Sol, HomeGPT 6.1 Sol
    Every model on TripIt
    ModelPassed%Screenshot
    GPT 6.1 Sol7215.1
    GLM 5.3 Flash6814.2
    Claude Opus 55311.1
    Grok 4.65311.1
    Kimi K35311.1
    Qwen 3.8 Max479.8—
    Claude Opus 5.5439.0
    DeepSeek V4.1 Flash387.9

    Every result in the WaffleRun the benchmark