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. Saturation forecast: Around January 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 0.43 | #98 |
| 2 | Gemini 2.5 Pro | 0.42 | #152 |
| 3 | GPT-5 Mini | 0.4 | #186 |
| 4 | Grok 4 | 0.39 | #181 |
| 5 | Gemini 2.5 Flash | 0.36 | #245 |
| 6 | Claude Opus 4 | 0.29 | #163 |
| 7 | Claude Sonnet 4 | 0.27 | #199 |
| 8 | GPT-4o | 0.13 | #343 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=blueprint-bench · How It Works · Data refreshed daily, snapshot 2026-10-20.