Concept art & fantasy
How 23 AI image models render: “A majestic white tiger with eyes like molten gold, walking through a misty, ancient forest. Strange, bioluminescent flora casts an ethereal glow on the gnarled trees and moss-covered stones. The atmosphere is one of profound mystery and quiet solitude. Rendered in the style of detailed matte painting with a touch of dark fantasy concept art.”
Same prompt, sent verbatim, at each model's default settings and a fixed seed. Judge: design originality, readable silhouette, lighting drama. 62 community votes so far.
Community prompt by @anyafoxx · voting closed 2026-10-05
23 text-to-image models rendered this concept art & fantasy prompt with the identical text at their default settings and a fixed seed. GPT Image 2 is the community favourite so far, winning 83% of its 6 head-to-head votes on this prompt. Grok Imagine Image Quality was the fastest at 5.6s.
What to look for: sense of scale, atmosphere and light, and designs that read clearly at a glance.
- GPT Image 283%OpenAI80.3s · 60 cr Try
- Muse Image 1.067%Meta12.3s · 12 cr Try
- Recraft v4 Pro60%Recraft · retired from voting21.9s · 300 cr Try
- 57%22.2s · 50 cr Try
- Krea 2 Medium57%Krea16.7s · 40 cr Try
- Qwen Image 3.057%Alibaba7.2s · 32 cr Try
- ImagineArt 1.5 Pro56%ImagineArt51.9s · 55 cr Try
- GPT Image 2.5 Flare50%OpenAI91.9s · 40 cr Try
- Ideogram v443%Ideogram15.9s · 40 cr Try
- GPT Image 2.5 Sunburst40%OpenAI82.3s · 40 cr Try
- Nano Banana 2 Lite33%Google14.9s · 36 cr Try
- Krea 2 Large29%Krea37.3s · 80 cr Try
- ImagineArt 225%ImagineArt30.6s · 60 cr Try
- 5.6s · 80 cr Try
- Kling v3 Omni17%Kuaishou96.0s · 35 cr Try
- Seedream 5.0 Pro17%ByteDance76.3s · 49 cr Try
- Qwen Image 3.0 ProAlibaba12.8s · 36 cr Try
- Seedream 5.0 FlashByteDance32.0s · 22 cr Try
- Recraft v4Recraft · retired from voting10.5s · 45 cr Try
- FLUX 3 ImageBlack Forest Labs32.1s · 32 cr Try
- Ideogram v4.5Ideogram15.2s · 80 cr Try
- Midjourney Niji 7Midjourney · retired from voting102.7s · 160 cr Try
FAQ
- Which AI model renders this prompt best?
- GPT Image 2 leads this prompt in the BudgetPixel AI Image Arena, winning 83% of its 6 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.