MazeBench 3D Spatial Reasoning: Opus 5.5 Hits 6% While GPT-6 Sol and Grok 4.7 Manage Just 1%
patience_cave · x · 2026-09-23
A tough new 3D spatial reasoning benchmark, MazeBench, humbles frontier models: Opus 5.5 leads with a 6% score, inching closer to Astra, while GPT-6 Sol and Grok 4.7 both score only 1%. The results underline how far top models remain from competent 3D spatial reasoning.
Related event: MazeBench Exposes Weak 3D Spatial Reasoning in Frontier Models(2 posts)→
More from Models
- DeepMind launches AlphaGenome 2 with major performance leap in variant-effect prediction — anshulkundaje · 2026-11-11
- US data vendors accused of selling RL environments to Chinese military-linked firms — xeophon · 2026-11-11
- Theory: Typos Disrupt LLM Tokenization, Which May Explain Why AI Text Has a Distinct Smell — menhguin · 2026-11-11
- LMArena co-founder explains AI evaluation methodology and policy in new podcast — MattPerault · 2026-11-11
- Runway open-sources Praxis-1, turning video pretraining into real robot control — runwayml · 2026-11-11
- DeepMind launches AlphaGenome 2, spotting nearly 2x more gene-regulating mutations — GoogleDeepMind · 2026-11-11