WebsiteBenchWebsiteBench

Apple

Rebuilt by Claude Opus 5.5: 304 of 700 hidden tests passed (43.4%)

reference.local/
Reference for Apple, HomeReference
rebuild.local/
Rebuild by Claude Opus 5.5, HomeClaude Opus 5.5
    Every model on Apple
    ModelPassed%Screenshot
    Claude Opus 5.530443.4
    Claude Opus 518826.9
    GPT 6.1 Sol16323.3
    GLM 5.3 Flash12017.1
    Grok 4.69613.7
    DeepSeek V4.1 Flash8011.4
    Kimi K3Timeout
    Qwen 3.8 MaxError—