AI Text to Image

Describe it in a sentence. Pick a model, a size, and how many you want.

Try:
Text to Image
Costs5credits

Sign in to start — new accounts get 20 free credits, no card.

Which model should I pick?

All of them work from the same prompt box — the price per image is the only thing you commit to. Start with the cheapest and move up if the result isn't sharp enough.

Nano Banana5 credits per image
Google Gemini 2.5 Flash · fastest, cheapest
FLUX.1 Kontext5 credits per image
Black Forest Labs · fastest, about 5s per image
Seedream 5 Lite7 credits per image
ByteDance · good at Chinese text and posters
Seedream 5 Pro · 2K18 credits per image
ByteDance · most photorealistic, highest detail
GPT Image 2 · 2K7 credits per image
OpenAI · good with text in images, half the price of 2.5
GPT Image 2 · 4K10 credits per image
OpenAI · for print or heavy cropping
GPT Image 2.5 · 2K13 credits per image
OpenAI · newest, better instruction following
GPT Image 2.5 · 4K20 credits per image
OpenAI · newest, highest resolution

Writing a prompt that works

Say what the subject is, then how it's lit, then the framing. "A ceramic mug on linen" gives you an average mug. "Matte black ceramic mug on pale linen, soft window light from the left, shallow depth of field" gives you the one you pictured. The three examples above the prompt box are built this way — click one and edit it.

Start from a photo instead

Already have an image and want to change it? Use Image to Image. To use Google's model specifically, the Nano Banana page preselects it for both modes.