IDEAL-Bench - Grid Coverage: leaderboard
Grid layout coverage on IDEAL-Bench (2026), a single-image 3D indoor layout benchmark: in classroom, library and office scenes, the percent of ground-truth grid cells that a predicted object of the target category falls into. Higher is better.
Metric: Grid coverage rate (%): share of ground-truth grid cells (classroom, library and office scenes) that a predicted object of the target category falls into, IDEAL-Bench 3D layout task; higher is better. Source: arxiv.org. Saturation forecast: Around September 2027. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 29.1 | #266 |
| 2 | GPT-5.4 | 28.6 | #123 |
| 3 | Claude Sonnet 4.6 | 27.8 | #161 |
| 4 | GPT-4o | 25.2 | #741 |
| 5 | Gemma 4 31B (IT) | 24.1 | #342 |
| 6 | Qwen 3 VL 30B A3B Instruct | 22.9 | #871 |
| 7 | Qwen 3 VL 235B A22B Instruct | 16.1 | #522 |
| 8 | Qwen 3 VL 8B Instruct | 14.4 | #981 |
| 9 | GLM-4.6V | 13.3 | #701 |
| 10 | Qwen 2.5 VL 72B Instruct | 11.9 | #846 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=ideal-bench-grid-coverage · How It Works · Data refreshed daily, snapshot 2026-11-13.