OpenAI upgrades ChatGPT Images to rival Google's Nano Banana: What's new

The updated ChatGPT Images, powered by GPT Image 1.5, is said to bring faster image generation, improved subject consistency and a new dedicated Images space within ChatGPT

OpenAI has announced a new version of ChatGPT Images, an updated image generation and editing model aimed at delivering faster output and more precise edits. Powered by what OpenAI calls its new flagship image model, GPT Image 1.5, the update is rolling out globally to ChatGPT users and in API for developers, as the company looks to strengthen its position against rival tools such as Google’s Nano Banana image editor.

 

According to OpenAI, the new ChatGPT Images experience is designed to improve both image creation from scratch and edits to existing photos, with a focus on preserving visual details across multiple changes.

What’s new in ChatGPT Images

OpenAI says the updated model is capable of generating images up to four times faster than the previous version, while adhering more closely to user instructions. When editing uploaded images, the model is designed to change only the elements specified by the user, while keeping aspects such as lighting, composition and people’s appearance consistent across edits. This is similar to what Google offers through its Gemini-based Nano Banana tool. 

The company claims this enables more usable results for tasks such as photo edits, clothing and hairstyle try-ons, stylistic filters and conceptual transformations. The model supports a range of editing operations, including adding, removing, blending, combining and transposing elements, without significantly degrading image quality.

 

Text rendering has also been improved, with the model now able to handle denser and smaller text more reliably, as per OpenAI. Additional quality improvements include better handling of multiple small faces and more natural-looking outputs overall.

 

Alongside the model update, OpenAI has introduced a dedicated Images space within ChatGPT, accessible via the sidebar on web and the mobile app. The new section includes preset styles, filters and prompts intended to make image generation and exploration faster, without requiring detailed written prompts.

The same improvements are also available through the API as GPT Image 1.5. OpenAI says the model offers better image preservation and brand consistency than its predecessor, making it suitable for use cases such as marketing assets, logo design and ecommerce product imagery.

Availability

The new ChatGPT Images model is rolling out globally starting today for all ChatGPT users and API customers, according to OpenAI. The feature works across models, meaning users do not need to manually select a specific option to access it. The earlier version of ChatGPT Images will continue to be available as a custom GPT.

 

OpenAI said the new Images experience within ChatGPT is also rolling out for most users today, with access for Business and Enterprise users planned for a later date.