DreamHouse (Planner-Atomic, Facade) - Joint Pass Rate: leaderboard
Metric: Joint pass rate (x 100): share of structures that both pass every structural test and reach visual fidelity 0.6 or more against the target views, on 1,200 DreamHouse residential timber-frame structures (13 architectural styles) per model and protocol, given a multi-view rendering of the finished exterior, which hides the framing, where the model writes one complete Blender Python construction script and regenerates it from validator feedback until all tests pass or retries run out; higher is better. Source: arxiv.org. Saturation forecast: Around 2030. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.5 | 6.4 | #79 |
| 2 | Gemini 3 Flash | 3.6 | #93 |
| 3 | GPT-5 | 1.9 | #91 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=dreamhouse-planner-atomic-facade-joint-pass-rate · How It Works · Data refreshed daily, snapshot 2026-10-11.