WebsiteBenchWebsiteBench

BoardGameGeek

Rebuilt by Claude Opus 5.5: 142 of 1,020 hidden tests passed (13.9%)

reference.local/
Reference for BoardGameGeek, HomeReference
rebuild.local/
Rebuild by Claude Opus 5.5, HomeClaude Opus 5.5
    Every model on BoardGameGeek
    ModelPassed%Screenshot
    Claude Opus 5.514213.9
    GPT 6.1 Sol12712.5
    Claude Opus 5676.6
    DeepSeek V4.1 Flash474.6
    Grok 4.6343.3
    GLM 5.3 Flash333.2
    Kimi K3292.8
    Qwen 3.8 Max90.9