Qwen-Image-Bench - Quality: leaderboard
Qwen-Image-Bench (2026) scores text-to-image models on 1,000 creator-designed prompts in five pillars: quality, aesthetics, alignment, real-world fidelity and creative generation. This board is the visual quality pillar score from the judge model on the Chinese prompts, 0 to 100; higher is better.
Metric: Visual quality pillar 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; higher is better. Source: arxiv.org. Saturation forecast: Rough model projection: around 2026. 18 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT Image 2 | 58.65 | |
| 2 | Nano Banana Pro | 55.67 | |
| 3 | GPT Image 1.5 | 55.14 | |
| 4 | Nano Banana 2 | 54.77 | |
| 5 | Seedream 4.5 | 54.41 | |
| 6 | Qwen Image 2.0 Pro | 54.39 | |
| 7 | Seedream 4.0 | 54.01 | |
| 8 | FLUX 2 Max | 53.64 | |
| 9 | Seedream 5.0 | 52.55 | |
| 10 | GPT Image 1 | 52.34 | |
| 11 | FLUX 2 Pro | 52.3 | |
| 12 | Qwen-Image-2512 | 51.76 | |
| 13 | Imagen 4.0 Ultra | 50.9 | |
| 14 | HunyuanImage 3.0 | 50.35 | |
| 15 | Imagen 4.0 | 50.16 |
Interactive version: theaggregate.ai/benchmark?slug=qwen-image-bench-quality · How It Works · Data refreshed daily, snapshot 2026-11-22.