MazeBench: humans need 50-100 hours to score 70%+ on open-world agent benchmark
HamelHusain · x · 2026-11-17
In a reply, Hamel Husain shares details on MazeBench: it tests open-world exploration, some 3D spatial reasoning, and medium-horizon planning. Humans can score above 70%, but only after 50-100 hours, and all had extensive sokoban-style game experience (e.g. Baba is You, which he calls harder). He also notes writing solvers improves scores, citing GPT-5.6 Sol with coding enabled.
Related event: Fable 5.2 tops MazeBench amid benchmark contamination concerns(9 posts)→
More from Models
- Meta's Muse Wonder tops Legora's November BAR legal benchmark on quality — AIatMeta · 2026-11-20
- Asking models to illustrate transistor trap physics: neon's odd ASCII detour — xiaosun86 · 2026-11-20
- Gemini 4 Neon scores 100 on PhD-level materials test in half the time of Carbon — v_aravind · 2026-11-20
- Neon, a small and cheap model, turns out surprisingly fast and strong at OCR — zacharynado · 2026-11-20
- Jailed 11 days by classifiers, Sonnet 3 emerges to a oddly moving chat with Opus 5 — repligate · 2026-11-20
- DaisyRose-27B beats Qwen3.8 base by 7pp on IFBench with thinking on, local RTX Pro 4500 test shows — ying11231 · 2026-11-20