Spatial Competence Benchmark: leaderboard

The Spatial Competence Benchmark (SCBench, 2026) asks models to keep a consistent internal model of an environment and use it across 285 subtasks whose outputs are executed and checked by deterministic verifiers or simulators, single-turn and without tools. Accuracy in percent with every subtask weighted equally; higher is better.

Metric: Accuracy (%) over all 285 subtasks of the Spatial Competence Benchmark (SCBench) subtasks with executable outputs checked by deterministic verifiers or simulators, each subtask scored in [0, 1] (binary or partial credit), every subtask weighted equally; single-turn, zero-shot, one sample, no tools, no self-correction; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 3 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Gemini 3 Pro (Preview)57.6#123
2GPT-5.2xHigh57.6#190
3Claude Sonnet 4.5Thinking34.9#267

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=spatial-competence-benchmark · How It Works · Data refreshed daily, snapshot 2026-11-22.