How 20 AI image models render: “A red bicycle leaning against a yellow wall, a black cat sitting on the bicycle's saddle, a small potted cactus to the left of the bicycle.”
Same prompt, sent verbatim, at each model's default settings and a fixed seed. Judge: every stated element present, correct count, correct placement. 126 community votes so far.
Core suite prompt · voting closed 2026-09-18
Checklist: red bicycle · yellow wall · black cat on the saddle · potted cactus · cactus to the LEFT of the bicycle
20 text-to-image models rendered this prompt adherence prompt with the identical text at their default settings and a fixed seed. Nano Banana 2 Lite is the community favourite so far, winning 70% of its 10 head-to-head votes on this prompt. Grok Imagine Image Quality was the fastest at 5.4s.
What to look for: the right count of objects, the right spatial relationships, the stated colours and attributes, and nothing added.
- Nano Banana 2 Lite70%Google25.9s · 36 cr Try
- Qwen Image 3.070%Alibaba7.4s · 32 cr Try
- Muse Image 1.067%Meta9.5s · 12 cr Try
- ImagineArt 267%ImagineArt · retired from voting29.6s · 60 cr Try
- Kling v3 Omni63%Kuaishou31.9s · 35 cr Try
- Seedream 5.0 Pro63%ByteDance29.6s · 49 cr Try
- GPT Image 2.5 Flare62%OpenAI26.4s · 40 cr Try
- GPT Image 262%OpenAI · retired from voting50.6s · 60 cr Try
- 5.4s · 80 cr Try
- ImagineArt 1.5 Pro55%ImagineArt · retired from voting52.9s · 55 cr Try
- GPT Image 2.5 Sunburst50%OpenAI24.5s · 80 cr Try
- Qwen Image 3.0 Pro43%Alibaba9.1s · 36 cr Try
- Recraft v4 Pro40%Recraft · retired from voting16.1s · 300 cr Try
- Ideogram v438%Ideogram13.4s · 40 cr Try
- Midjourney Niji 725%Midjourney · retired from voting102.4s · 160 cr Try
- Krea 2 Medium20%Krea17.4s · 40 cr Try
- 20%14.5s · 50 cr Try
- Recraft v418%Recraft · retired from voting10.8s · 45 cr Try
- Krea 2 Large0%Krea36.8s · 80 cr Try
FAQ
- Which AI model renders this prompt best?
- Nano Banana 2 Lite leads this prompt in the BudgetPixel AI Image Arena, winning 70% of its 10 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.