Qwen-Image-Bench: leaderboard
Metric: Overall Q-Judger score (0-100), the unified judge model fine-tuned from Qwen3.6-27B on expert annotations, aggregated bottom-up from 56 third-level facets, over the 1,000 creator-designed Chinese prompts (500 short, 500 long); higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT Image 2 | 64.69 |
| 2 | Nano Banana 2 | 59.82 |
| 3 | GPT Image 1.5 | 59.65 |
| 4 | Nano Banana Pro | 59.45 |
| 5 | Qwen Image 2.0 Pro | 57.84 |
| 6 | Seedream 5.0 | 57.22 |
| 7 | Seedream 4.5 | 56.78 |
| 8 | Seedream 4.0 | 56.21 |
| 9 | FLUX 2 Max | 55.33 |
| 10 | FLUX 2 Pro | 54.57 |
Interactive version: theaggregate.ai/benchmark?slug=qwen-image-bench · How It Works · Data refreshed daily, snapshot 2026-10-07.