IDEAL-Bench: leaderboard
Metric: Overall score (0-100): unweighted mean of the 11 rate metrics on 1,000 photorealistic indoor scenes across 10 room types, where a VLM predicts the room type, room size and every visible object's category, 3D position, size and yaw as JSON from one image (zero-shot, temperature 0.1, no chain-of-thought); higher is better. Source: arxiv.org. Saturation forecast: Around June 2028. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 62.1 |
| 2 | GPT-4o | 60.4 |
| 3 | Gemma 4 31B (IT) | 59.7 |
| 4 | Claude Sonnet 4.6 | 59.2 |
| 5 | GPT-5.4 | 56.7 |
| 6 | Qwen 3 VL 8B Instruct | 55.1 |
| 7 | GLM-4.6V | 54.4 |
| 8 | Qwen 2.5 VL 72B Instruct | 52.6 |
| 9 | Qwen 3 VL 235B A22B Instruct | 51.8 |
| 10 | Qwen 3 VL 30B A3B Instruct | 50.8 |
| 11 | InternVL3.5-8B | 45.9 |
Interactive version: theaggregate.ai/benchmark?slug=ideal-bench · How It Works · Data refreshed daily, snapshot 2026-09-29.