Typography
How 22 AI image models render: “Within a quiet mineral gallery, amethyst clusters and polished geodes catch pools of light; "PALE ATLAS" is carved in clean capitals across a slate plaque beside the entrance.”
Same prompt, sent verbatim, at each model's default settings and a fixed seed. Judge: is the text exactly right, spelled correctly, well set? 122 community votes so far.
Fresh prompt from the daily stream · voting closed 2026-10-08
Expected text: PALE ATLAS
22 text-to-image models rendered this typography prompt with the identical text at their default settings and a fixed seed. Qwen Image 3.0 Pro is the community favourite so far, winning 73% of its 11 head-to-head votes on this prompt. Qwen Image 3.0 was the fastest at 7.0s.
What to look for: correct spelling, placement and legibility, and type that looks printed or carved rather than pasted on.
- Qwen Image 3.0 Pro73%Alibaba20.8s · 36 cr Try
- 7.5s · 80 cr Try
- Recraft v4 Pro67%Recraft · retired from voting13.0s · 300 cr Try
- GPT Image 2.5 Sunburst64%OpenAI75.2s · 40 cr Try
- GPT Image 2.5 Flare57%OpenAI55.1s · 40 cr Try
- Ideogram v4.554%Ideogram15.5s · 80 cr Try
- 53%16.3s · 50 cr Try
- Muse Image 1.050%Meta9.0s · 12 cr Try
- Nano Banana 2 Lite50%Google21.2s · 36 cr Try
- Kling v3 Omni45%Kuaishou44.1s · 35 cr Try
- GPT Image 242%OpenAI63.4s · 42 cr Try
- Qwen Image 3.042%Alibaba7.0s · 32 cr Try
- ImagineArt 233%ImagineArt27.3s · 60 cr Try
- Seedream 5.0 Pro33%ByteDance34.7s · 49 cr Try
- Ideogram v433%Ideogram14.8s · 40 cr Try
- Krea 2 Large31%Krea85.0s · 80 cr Try
- Seedream 5.0 Flash25%ByteDance25.1s · 22 cr Try
- ImagineArt 1.5 Pro10%ImagineArt · retired from voting49.1s · 55 cr Try
- Krea 2 Medium0%Krea36.6s · 40 cr Try
- Nano Banana 2.1Google9.8s · 40 cr Try
- FLUX 3 ImageBlack Forest Labs37.3s · 32 cr Try
FAQ
- Which AI model renders this prompt best?
- Qwen Image 3.0 Pro leads this prompt in the BudgetPixel AI Image Arena, winning 73% 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.