WebsiteBenchWebsiteBench

H Mart

Rebuilt by Claude Opus 5.5: 331 of 714 hidden tests passed (46.4%)

reference.local/
Reference for H Mart, HomeReference
rebuild.local/
Rebuild by Claude Opus 5.5, HomeClaude Opus 5.5
    Every model on H Mart
    ModelPassed%Screenshot
    Claude Opus 5.533146.4
    DeepSeek V4.1 Flash29341.0
    Claude Opus 525035.0
    GLM 5.3 Flash24334.0
    GPT 6.1 Sol24133.8
    Grok 4.623733.2
    Kimi K320829.1
    Qwen 3.8 Max00.0—