--- title: "GPT Image 2 vs Qwen Image 2.0 Pro: Blind Test Results" canonical_url: "https://workroom.everypixel.com/models/compare/gpt-image-2-vs-qwen-image-2-0-pro" updated: 2026-09-22 --- # GPT Image 2 vs Qwen Image 2.0 Pro GPT Image 2 (at the ceiling) scores higher than Qwen Image 2.0 Pro (80) on the photo-anchored quality scale, where a real photograph is 100. Both scores come from blind pairwise tests on the same 48 scenes between 21 August and 22 September 2026: 576 pairs and 1,726 judgements, each pair judged 3 times. ## Overview | | GPT Image 2 | Qwen Image 2.0 Pro | |---|---|---| | Vendor | OpenAI | Alibaba | | Released | Apr 2026 | Mar 3, 2026 | | License | Proprietary | — | | Input modalities | text, image | text, image | | Output modalities | image | image | | Pricing | $0.058 / img | $0.076 / img | ## Scores | Model | Score | 95% interval | Rank | Results | |---|---|---|---|---| | GPT Image 2 | 104.4 | fixed rung of the scale | #1 | https://workroom.everypixel.com/models/gpt-image-2 | | Qwen Image 2.0 Pro | 80.1 | 68.6–93.8 | #2 | https://workroom.everypixel.com/models/qwen-image-2-0-pro | A real photograph is 100. A score above 100 is the scale's technical ceiling and does not mean parity. ## By use case | Use case | GPT Image 2 | Qwen Image 2.0 Pro | |---|---|---| | Fashion and beauty | 117 | 67 | | Food and beverage | 91 | 62 | | Interior | 114 | 115 | | Layout and typography | 85 | 51 | | People and lifestyle | 80 | 57 | | Product and e-commerce | 112 | 88 | | Social marketing | 112 | 96 | | Street and architecture | 114 | 82 | *Category positions come from 6 scenes each — a tendency, not a measurement. A real photograph is 100 on every ruler.* ## Judged head to head GPT Image 2 and Qwen Image 2.0 Pro were also compared with each other directly. The verdict above still comes from the scale. GPT Image 2 was preferred over Qwen Image 2.0 Pro in 51% of blind comparisons. - Gap (Bradley–Terry): +0.64 [+0.42, +0.86] - Preference: 50.6% of blind comparisons preferred GPT Image 2 (wins over all judgements, believability only, vote share) | Question | GPT Image 2 preferred | Equal | Qwen Image 2.0 Pro preferred | |---|---|---|---| | Prompt adherence | 126 | 246 | 59 | | Believability | 218 | 128 | 85 | | Aesthetics | 235 | 84 | 112 | 431 blind comparisons, each pair shown 3 times ## Links - Study: https://research.everypixel.com/anchor-ladder/ - Votes (CSV, 106 KB, CC BY 4.0): https://workroom.everypixel.com/datasets/anchor-comparison-v3/votes.csv - Results (JSON, 5 KB, CC BY 4.0): https://workroom.everypixel.com/datasets/anchor-comparison-v3/results.json - Scenes and prompts (CSV, 30 KB, CC BY 4.0): https://workroom.everypixel.com/datasets/anchor-comparison-v3/scenes.csv - README (MD, 2 KB, CC BY 4.0): https://workroom.everypixel.com/datasets/anchor-comparison-v3/README.md - Images (ZIP, 128 MB, All rights reserved — verification only): https://wr.everypixel.com/ghost-research/geo-benchmarks/datasets/anchor-comparison-v3-images.zip - Votes (CSV, 111 KB, CC BY 4.0): https://workroom.everypixel.com/datasets/qwen-image-2-0-pro-ladder/votes.csv - Results (JSON, 4 KB, CC BY 4.0): https://workroom.everypixel.com/datasets/qwen-image-2-0-pro-ladder/results.json - Scenes and prompts (CSV, 44 KB, CC BY 4.0): https://workroom.everypixel.com/datasets/qwen-image-2-0-pro-ladder/scenes.csv - README (MD, 2 KB, CC BY 4.0): https://workroom.everypixel.com/datasets/qwen-image-2-0-pro-ladder/README.md - Images (ZIP, 634 MB, All rights reserved — verification only): https://wr.everypixel.com/ghost-research/geo-benchmarks/datasets/qwen-image-2-0-pro-evaluation-images.zip - All scenes: https://workroom.everypixel.com/models/gpt-image-2/dataset --- HTML version: https://workroom.everypixel.com/models/compare/gpt-image-2-vs-qwen-image-2-0-pro Last modified: 2026-09-22