New from OpenAI

GPT Image 2.5: posters, ads and photo edits

GPT Image 2.5 is OpenAI's image model, released on September 8, 2026. It is strongest where many image models fail: posters and ads with a lot of text, a clean layout and correct lettering, including Cyrillic. It also makes photoreal portraits and edits the photos you attach. On avots.ai it is the model our router picks for posters and ads, and you pay per picture.

Try GPT Image 2.5 now ✦ See pricing
Specs

GPT Image 2.5 at a glance

Text that reads right

Headlines, prices and buttons come out spelled correctly; in our test poster the Russian text had no mistakes.

Layouts, not just pictures

It builds a whole ad: headline, offer, icons and a call to action in one image.

Up to 10 images per request

Ask for ten variants and they come back from one request.

Photo edits

Attach up to 4 photos and describe the change; the picture keeps its shape.

About 10 to 17 seconds

The time per image we measured on avots.ai.

About 40 tokens a picture

On the standard price; the app shows the exact cost before you create.

Compare

GPT Image 2.5 next to Nano Banana and Recraft

 GPT Image 2.5Nano Banana 2Recraft 4.1 Flash
Best atposters and ads with textphoto edits that keep the facefast illustration and design
Speed per imageabout 10 to 17 sabout 10 sabout 10 s
Pictures per requestup to 10up to 10one
Edits attached photosyes, up to 4yesno
On avots.ailive nowlive nowlive now

From our own side-by-side test on the same prompts in September 2026. On avots.ai every image model runs from one balance, so you can try one prompt on all of them.

Good at

What GPT Image 2.5 is good at

Posters and banners

Event posters, sale banners and social posts with a headline that reads cleanly.

Ads for a small business

A café, a shop or a service: the product, the offer and a button in one picture.

Photoreal portraits

Natural light, real skin and believable settings.

Edits by description

Change clothes, background or details on a photo you upload.

Make a poster with GPT Image 2.5
Three steps

How to use GPT Image 2.5

Create an account

Sign up on app.avots.ai with email, Google or Telegram. Takes a minute.

Top up from 0.99 EUR

1,000 tokens cost 0.99 EUR. No plan to choose, no renewal, the balance sits there until you use it.

Pick GPT Image 2.5

Choose it in the image picker, or just ask Agent Avots for a poster or an ad and it picks GPT Image 2.5 for you.

One balance

Where you can use GPT Image 2.5

Web app

app.avots.ai on desktop and mobile, no install needed.

Telegram

GPT Image 2.5 in @AvotsAIbot, same balance, from any phone.

MCP

Use it inside Claude Desktop, Cursor or Cline via the MCP server.

Make your first poster in a minute

New accounts get welcome tokens, enough for several GPT Image 2.5 pictures. The same balance covers Nano Banana, Recraft, FLUX and video models, with no subscription.

Sign up and create ✦
More models

Related AI models

Nano Banana

Google Gemini image model for photoreal images and edits.

Recraft

Recraft 4.1 Flash for fast illustration and Recraft V3 for vector.

GPT-5 Image

The previous OpenAI image models, still available.

FLUX

FLUX 1.1 Pro and FLUX.2 for photoreal images.

Seedream 4.5

ByteDance image model for rich, detailed scenes.

avots.ai runs 40+ AI models for chat, image, video and audio on one balance. See the full list on the Tools page or open the app and start creating.

Browse all model pages Open the app
FAQ

GPT Image 2.5 questions

What is GPT Image 2.5?
GPT Image 2.5 is OpenAI's image model, released on September 8, 2026. It is strongest at posters and ads with a lot of text and a clean layout, makes photoreal portraits, and edits photos you attach by description.
How much does GPT Image 2.5 cost on avots.ai?
About 40 tokens a picture at the standard price, pay per use with no subscription. 1,000 tokens cost 0.99 EUR, and the app shows the exact cost before anything is charged. Editing an attached photo costs about twice as much.
Can GPT Image 2.5 write text in images?
Yes, that is its strongest side. Headlines, prices and button labels come out spelled correctly, and in our test the Russian text on a poster had no mistakes.
How many images can I make at once?
Up to 10 pictures from one request. Write the number in your message, for example «make 5 variants», and they arrive together.
GPT Image 2.5 vs Nano Banana, which to use?
GPT Image 2.5 for posters, ads and anything with text. Nano Banana for editing a photo of a person when the face must stay exactly the same. On avots.ai both are on one balance, and the agent picks between them for you.
Try it on avots ✦
Start creating