WebsiteBenchWebsiteBench

Bean Box

Rebuilt by Claude Opus 5.5: 322 of 987 hidden tests passed (32.6%)

reference.local/
Reference for Bean Box, HomeReference
rebuild.local/
Rebuild by Claude Opus 5.5, HomeClaude Opus 5.5
    Every model on Bean Box
    ModelPassed%Screenshot
    Claude Opus 5.532232.6
    Claude Opus 528428.8
    DeepSeek V4.1 Flash27027.4
    Kimi K326526.8
    Grok 4.623423.7
    Qwen 3.8 Max404.1
    GLM 5.3 Flash383.9—
    GPT 6.1 Sol90.9