Image generation with reliable text rendering — the only model that puts readable text inside images
GPT Image creates and edits images using OpenAI's model via fal.ai. Its standout capability: it reliably renders readable text inside images — recipe cards, menus, infographics, labels, certificates. No other image model does this consistently. Also offers precise image editing with quality tiers (low/medium/high).
Generate recipe cards, restaurant menus, and food labels with readable text and appealing visuals
Create infographics with data labels, headings, and explanatory text baked into the image
Generate certificates, awards, and badges with custom names and text
Make targeted edits to images — change specific objects, fix details, adjust elements
Say "use gpt image" or "generate with gpt image". Without this, the assistant defaults to Nano Banana.
Use GPT Image when you need text inside the image or precise edits.
Three tiers: low (fastest, cheapest), medium (default), high (most detail). Say "high quality" to upgrade.
Send an image and describe the edit. GPT Image excels at targeted changes like "change the label text to..." or "replace the logo with..."
Use gpt image to create a recipe card for pasta carbonara, high quality
MEDIA:pasta-carbonara-recipe.png Generated with GPT Image (high quality). Ingredients and instructions are readable. Want me to adjust the layout?
Edit this menu image — change the price of the Caesar salad to $14.99
MEDIA:menu-edited.png Price updated. The font matches the original menu style.
Based on our observed average over the last 4 weeks of real usage (not vendor list price). Generation and editing have different per-call costs; quality tier (low/medium/high) does not affect provider cost.
LLM processing cost is additional and depends on conversation complexity. BYOK users pay LLM costs directly to their provider.
* Prices include platform service fee. Actual costs may vary.