Photorealism
How 22 AI image models render: “Candid documentary photograph of an adult waiting beside a city bus shelter at dawn, sodium streetlights fading into cool blue daylight, 50mm lens with shallow depth of field, fine rain droplets clinging to the shelter glass.”
Same prompt, sent verbatim, at each model's default settings and a fixed seed. Judge: could this be a photograph? Skin, light, lens behaviour. 109 community votes so far.
Fresh prompt from the daily stream · voting closed 2026-10-08
22 text-to-image models rendered this photorealism prompt with the identical text at their default settings and a fixed seed. GPT Image 2.5 Flare is the community favourite so far, winning 91% of its 11 head-to-head votes on this prompt. Grok Imagine Image Quality was the fastest at 6.5s.
What to look for: believable skin and fabric, physically plausible light and shadow, lens behaviour such as depth of field and motion blur, and no waxy or over-smoothed surfaces.
- GPT Image 2.5 Flare91%OpenAI28.6s · 40 cr Try
- Muse Image 1.064%Meta12.2s · 12 cr Try
- 6.5s · 80 cr Try
- ImagineArt 1.5 Pro57%ImagineArt · retired from voting51.2s · 55 cr Try
- Krea 2 Large56%Krea86.7s · 80 cr Try
- Ideogram v455%Ideogram15.0s · 40 cr Try
- Qwen Image 3.055%Alibaba7.2s · 32 cr Try
- Recraft v4 Pro50%Recraft · retired from voting14.4s · 300 cr Try
- GPT Image 246%OpenAI60.0s · 42 cr Try
- Nano Banana 2 Lite44%Google22.0s · 36 cr Try
- GPT Image 2.5 Sunburst43%OpenAI45.1s · 40 cr Try
- Ideogram v4.538%Ideogram13.0s · 80 cr Try
- Seedream 5.0 Pro38%ByteDance26.5s · 49 cr Try
- Kling v3 Omni30%Kuaishou35.5s · 35 cr Try
- Seedream 5.0 Flash29%ByteDance33.0s · 22 cr Try
- 23%14.4s · 50 cr Try
- FLUX 3 Image20%Black Forest Labs31.8s · 32 cr Try
- ImagineArt 211%ImagineArt27.2s · 60 cr Try
- Krea 2 Medium0%Krea34.6s · 40 cr Try
- Qwen Image 3.0 Pro0%Alibaba10.2s · 36 cr Try
- Nano Banana 2.1Google10.4s · 40 cr Try
FAQ
- Which AI model renders this prompt best?
- GPT Image 2.5 Flare leads this prompt in the BudgetPixel AI Image Arena, winning 91% of its 11 blind head-to-head votes. Vote on the arena page to add your judgement.
- Were the prompts adjusted per model?
- No. Every model received the exact same prompt, unmodified, at its default settings, with no negative prompt or style suffix. Some models rewrite prompts internally, and that is part of what is being compared.
- Can I run this prompt myself?
- Yes. Each tile’s “Try” link opens the BudgetPixel workshop with this prompt and that model preselected.