Blueprint-Bench — leaderboard

Tests spatial reasoning by asking models to convert ~20 apartment photographs into accurate 2D floor plans. Most models perform at or below the random baseline (0.279); humans score 0.547.

Metric: Similarity Score. Source: andonlabs.com. Status: saturation imminent. 12 models tracked.

Top models

#ModelScore
1GPT-50.43
2Gemini 2.5 Pro0.42
3GPT-5 Mini0.4
4Grok 40.39
5Gemini 2.5 Flash0.36
6Claude Opus 40.29
7Claude Sonnet 40.27
8GPT-4o0.13

Interactive version: theaggregate.ai/benchmark?slug=blueprint-bench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.