MazeBench results: flagship models score 0% on maze tasks
patience_cave · x · 2026-08-24
New MazeBench results are in: ox-alpha, Grok 4.6, GLM-5.3, and Qwen 3.8-max all score 0% on maze-solving tasks; Gemini 3.7 Flash manages 1% and beats Kimi K3. The takeaway: frontier models remain collectively terrible at structured spatial reasoning.
Related event: MazeBench maze eval: flagship models score 0%, Gemini 3.7 Flash leads(4 posts)→
More from Models
- Unredacted Anthropic filing exposes internal data-quality debate months before Opus 3 cutoff — zacharynado · 2026-10-15
- Google says teams across the company are using Gemini 4 Argon on its hardest engineering problems — CurieuxExplorer · 2026-10-15
- Perplexity's new embedding models: index with a 9B model, search with 0.6B — antoine_chaffin · 2026-10-15
- 797GB of n-gram counts in CPU RAM let 1B models match 57% more compute — PangWeiKoh · 2026-10-15
- Leak: Moonshot's Kimi flagship 'x4' done with pre-training, K3.1 in internal testing — lyraxana · 2026-10-15
- Unsloth ships GGUF quants for Nemotron 3.5: Bolt-74B runs in 24GB RAM — danielhanchen · 2026-10-15