After teasing a new AI image generation model, OpenAI has finally launched ChatGPT Images, a dedicated visual workspace powered by their new flagship model, GPT Image 1.5. For a really long time, Google's Nano banana was considered one of the best image generation tools launched in 2025. Now it is time to see if it holds up against this new AI image generation model by OpenAI.
OpenAI says the new ChatGPT Images offers stronger instruction following, precise editing, detail preservation, and is 4x faster than the previous ChatGPT image generation model.
Notion AI: Notion's AI features can enhance your workflow by helping you write, summarize, brainstorm, and automate tasks with simple prompts.
What's new (and what it actually means)
At a high level, this release is about control, getting the change you asked for, and keeping everything else stable. OpenAI highlights:
- Stronger instruction following which causes less creative guessing and more literal execution.
- Precise editing that focuses on your request instead of rewriting the whole image.
- Detail preservation, including maintaining lighting, composition, and people's appearances across edits.
- Up to 4× faster generation, so iteration feels less like waiting in line.
Editing that changes only what you asked for
The most important claim here is consistency across iterations. OpenAI claims it can change only what you ask for while keeping important elements, like lighting, composition, and people's appearance, the same in all inputs, outputs, and follow-up edits.
It also supports a wider range of edit operations, which matters if you're doing real creative assembly rather than one-off prompts. OpenAI specifically calls out editing that includes adding, subtracting, combining, blending, and transposing, the kinds of actions you'd normally do in layers in a design tool.
Better text rendering
Text has always been a weak spot for AI image models. OpenAI says this version takes another step ahead in text rendering, handling denser and smaller text more reliably. That is a big deal for things like posters, packaging drafts, menus, simple UI mockups, and infographic-style layouts, places where almost correct text is basically unusable.
A new Images space inside ChatGPT
OpenAI is also shipping a dedicated Images area in the ChatGPT sidebar (mobile and web) meant to make image generation feel less buried inside a chat thread. It includes preset filters and trending prompts to kickstart ideas.
For developers: GPT Image 1.5 in the API
In terms of API, OpenAI says GPT Image 1.5 offers stronger preservation and editing than the prior GPT Image 1. They also claim it maintains logos and key visuals more consistently during edits. This makes the new image model suitable for branding, marketing, and e-commerce catalogs.
Two business-relevant notes:
- OpenAI says image inputs and outputs are 20% cheaper in GPT Image 1.5 compared to GPT Image 1.
- The cookbook guide shows the model name as gpt-image-1.5 and frames it as suited for iterative, controllable creative production.
In Conclusion:
After a long time, the AI image generation battle is back with Google's Nano banana and ChatGPT's GPT Image 1.5. While both are very capable, only a side-by-side comparison will show which model actually offers better instruction-following, precise editing, and detail preservation while remaining speedy. ChatGPT Images is rolling out in ChatGPT for all users, and in the API as GPT Image 1.5.
💡 For Partnership/Promotion on AI Tools Club, please check out our partnership page.
