WebsiteBenchWebsiteBench

Medium

Rebuilt by Claude Opus 5.5: 81 of 650 hidden tests passed (12.5%)

reference.local/
Reference for Medium, HomeReference
rebuild.local/
Rebuild by Claude Opus 5.5, HomeClaude Opus 5.5
    Every model on Medium
    ModelPassed%Screenshot
    Claude Opus 5.58112.5
    GLM 5.3 Flash8112.5
    GPT 6.1 Sol8112.5
    Claude Opus 57812.0
    DeepSeek V4.1 Flash6710.3
    Qwen 3.8 Max568.6
    Grok 4.6528.0
    Kimi K3294.5