Skip to content
logo The magazine for digital lifestyle and entertainment
Artificial intelligence CoBi News All topics
ChatGPT Images 2.0

Better AI Images? OpenAI Expands Features With Update

I'm sorry, but I need the actual text of the article snippet to provide a translation. Could you please provide the text you would like translated?
With ChatGPT Images 2.0, AI-generated images are expected to be created more precisely according to specifications. Photo: Getty Images
Share article

April 25, 2026, 4:38 am | Read time: 2 minutes

Anyone who creates AI images knows the problem: The more precise the idea, the more difficult the implementation often becomes. OpenAI aims to address this issue.

Often, more than a brief description is needed to create an image according to one’s vision. Motifs should be in the right place, text must be legible, and the format should suit the intended use. To address this, OpenAI is now offering ChatGPT Images 2.0. The new image model is available for ChatGPT, Codex, and via an interface.

More Precise Results with Specific Instructions

According to the developers, Images 2.0 is better at understanding detailed instructions than older versions. This is especially true for prompts that contain multiple requests simultaneously. The AI is expected to reproduce text in images more accurately and place elements more reliably where they belong. OpenAI also promises more consistent results for diagrams, user interfaces, and more complex layouts.

In terms of format, Images 2.0 offers more flexibility. The model covers aspect ratios between 3:1 and 1:3, allowing for both very wide and portrait-oriented motifs. OpenAI cites banners and mobile formats as examples. For users, this means an image can be more closely tailored to its future use without being restricted to a standard size.

More on the topic

Image Content in Multiple Languages

Another aspect is language support. Images 2.0 can generate content in multiple languages. This can be important when not only a motif but also words or short texts are to appear directly in the image. OpenAI sees this as an advantage for international applications.

In ChatGPT, Images 2.0 can also work with so-called “Thinking” workflows, where a task is processed step by step. Additionally, multiple results can be output. Up to eight image variants are possible per request, allowing users to compare different drafts before deciding on a version.

Also of interest: Image generation with ChatGPT allows for the forgery of receipts and documents

Availability and API for Custom Applications

The model is now available to all users of ChatGPT and Codex. However, some advanced features depend on paid subscriptions. Those who want to use Images 2.0 not only directly in ChatGPT can also integrate it into their own applications via the “gpt-image-2” API.

This article is a machine translation of the original German version of TECHBOOK and has been reviewed for accuracy and quality by a native speaker. For feedback, please contact us at info@techbook.de.

You have successfully withdrawn your consent to the processing of personal data through tracking and advertising when using this website. You can now consent to data processing again or object to legitimate interests.