The best image generation models, according to hackers
22 image generation models were recommended on Hacker News between November 2025 and November 2026. We rank them by how many separate people recommend each one, and 4 earn the Hacker Approved badge.
Top picks
The three most recommended image generation models with the badge. Each figure is 5 people. RecommendsCriticizes One comment
Follows prompts faithfully
Praised for faithful prompt adherence, photorealism and clean text, and for reasoning before rendering. Critics mostly flag strict censorship, a yellow tint and drift over multiple edits.
Fast on consumer GPUs
People recommend it for fast, realistic local image generation with good prompt adherence and legible English and Chinese text. Critics say it struggles outside photographic portraits and with complex prompts.
Fast generation for its size
Open image generation model that runs locally, producing fast, photorealistic results with strong prompt adherence and editing. Caveat: imperfect text and prompt following.
All 22 image generation models, ranked
Select a column heading to sort. Approval share is recommenders out of everyone who gave an opinion; the tick on each bar marks the 2 to 1 line the badge needs. Months counts the months, out of 13, in which someone recommended it.
| Rank | |||||||
|---|---|---|---|---|---|---|---|
| 1 | Nano Banana | 265 | 159 | 63% | 12 | 6 mo5 owners | — |
| 2 | gpt-image-1 | 79 | 79 | 50% | 12 | — | — |
| 3 | FLUX.2 | 66 | 51 | 56% | 10 | — | — |
| 4 | gpt-image-2 | 63 | 26 | 71% | 8 | — | Approved |
| 5 | Qwen-Image | 47 | 38 | 55% | 10 | — | — |
| 6 | Midjourney | 37 | 20 | 65% | 11 | — | — |
| Tied 7 | Z-Image Turbo | 29 | 2 | 94% | 9 | — | Approved |
| Tied 7 | Z-Image | 29 | 7 | 81% | 7 | — | Approved |
| 9 | Stable Diffusion | 27 | 17 | 61% | 12 | — | — |
| Tied 10 | Krea 3 | 25 | 23 | 52% | 6 | — | — |
| Tied 10 | OpenAI image generation | 25 | 23 | 52% | 6 | — | — |
| 12 | Grok Imagine | 16 | 22 | 42% | 10 | — | — |
| 13 | Gemini 3 Pro Image | 15 | 11 | 58% | 8 | — | — |
| Tied 14 | Qwen-Image-Edit | 13 | 1 | 93% | 6 | — | Approved |
| Tied 14 | Seedream | 13 | 10 | 57% | 5 | — | — |
| 16 | DALL-E | 12 | 9 | 57% | 5 | — | — |
| 17 | Moebius image inpainting model | 9 | 7 | 56% | 2 | — | — |
| Tied 18 | Stable Diffusion SDXL | 8 | 4 | 67% | 5 | — | — |
| Tied 18 | Ideogram | 8 | 1 | 89% | 5 | — | — |
| 20 | Recraft | 7 | 1 | 88% | 5 | — | — |
| 21 | Retro Diffusion | 6 | 0 | 100% | 4 | — | — |
| 22 | Flux Kontext | 5 | 1 | 83% | 4 | — | — |
How we rank
Rank is the number of separate people who recommend a product in a Hacker News comment over the past year. Someone who praises the same image generation model in five threads counts once.
The badge goes to image generation models that at least 10 people recommend, with recommenders outnumbering critics at least 2 to 1, recommended in at least 4 of the 13 months.
Rankings are never for sale. We may earn a commission when you buy through our links. It never changes the order.