Blueprint-Bench 2: leaderboard
Spatial-reasoning benchmark measuring how accurately models convert apartment photos into 2D floor plans.
Metric: Connectivity Similarity Score. Source: andonlabs.com. Source: www-cdn.anthropic.com. Source: epoch.ai. Saturation forecast: Around May 2027. 33 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 4 Argon | 0.54 | #8 |
| 2 | Claude Opus 5.5 | 0.51 | #3 |
| 3 | GPT-6 Astra | 0.5 | #6 |
| 4 | Claude Fable 5.1 | 0.42 | #9 |
| 5 | Claude Sonnet 5.5 | 0.39 | #11 |
| 6 | Gemini 3.8 Flash | 0.39 | #21 |
| 7 | Claude Fable 5 | 0.39 | #16 |
| 8 | Claude Mythos 5 | 0.39 | #2 |
| 9 | GPT-6 Sol | 0.37 | #14 |
| 10 | GPT-5.5 | 0.36 | #31 |
| 11 | GPT-5.6 Sol | 0.34 | #19 |
| 12 | Gemini 3.5 Flash | 0.34 | #60 |
| 13 | Grok 4.6 | 0.33 | #26 |
| 14 | Claude Haiku 5.5 | 0.33 | #40 |
| 15 | Grok 4.7 | 0.33 | #25 |
No result here: #7 GPT-6.1 Sol, #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=blueprint-bench-2 · How It Works · Data refreshed daily, snapshot 2026-10-20.