Qwen-Image-Bench - Alignment: 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 prompt alignment pillar score from the judge model on the Chinese prompts, 0 to 100; higher is better.
Metric: Prompt alignment 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 | 65.85 | |
| 2 | Nano Banana 2 | 62.4 | |
| 3 | GPT Image 1.5 | 61.72 | |
| 4 | Nano Banana Pro | 61.25 | |
| 5 | Qwen Image 2.0 Pro | 59.28 | |
| 6 | Seedream 5.0 | 58.9 | |
| 7 | FLUX 2 Max | 57.35 | |
| 8 | Seedream 4.5 | 57.31 | |
| 9 | FLUX 2 Pro | 57.01 | |
| 10 | Seedream 4.0 | 56.64 | |
| 11 | GPT Image 1 | 56.28 | |
| 12 | Imagen 4.0 Ultra | 54.02 | |
| 13 | Qwen-Image-2512 | 52.72 | |
| 14 | HunyuanImage 3.0 | 52 | |
| 15 | Imagen 4.0 | 51.64 |
Interactive version: theaggregate.ai/benchmark?slug=qwen-image-bench-alignment · How It Works · Data refreshed daily, snapshot 2026-11-21.