OpenAI has released a new version of ChatGPT Images, powered by its latest flagship image generation model. The update brings faster image generation, improved instruction-following, and more precise editing capabilities, addressing long-standing challenges in AI-generated visuals such as maintaining consistency and accurately applying complex changes.

According to OpenAI, the rollout begins today for all ChatGPT users and API customers worldwide, with access for Business and Enterprise users expected to follow. The update reflects OpenAI’s continued focus on making image generation more reliable for both everyday users and professional workflows.

The model is also available via the API as GPT Image 1.5, replacing earlier image-generation capabilities. OpenAI says the updated model is designed to support a wide range of use cases—from creative experimentation to commercial and enterprise applications—by offering greater control, speed, and fidelity in image outputs.

OpenAI says one of the most significant improvements in the new ChatGPT Images model is its ability to apply precise edits while preserving key visual details. When users upload an image and request changes, the model now follows instructions more reliably, modifying only the specified elements while maintaining consistency in areas such as lighting, composition, and people’s appearance across multiple edits.

According to OpenAI, this increased precision enables more realistic photo edits and more believable clothing and hairstyle try-ons. These improvements are intended to reduce visual drift when users make iterative edits to the same image.

The model supports a wide range of editing actions, including adding, removing, blending, combining, and transposing elements. OpenAI says these capabilities allow users to make practical adjustments without degrading overall image quality or losing important visual context.

OpenAI says the updated ChatGPT Images model supports creative transformations that change or add visual elements—such as text and layout—while preserving important details from the original image. These capabilities are designed to help users bring more conceptual or expressive ideas to life without losing visual consistency.

According to OpenAI, these transformations work for both simple and more intricate concepts, allowing users to experiment with a wide range of visual styles and compositions. The model is intended to maintain core characteristics such as lighting, structure, and subject appearance, even as new elements are introduced.

OpenAI also notes that many of these creative transformations can be explored using preset styles and built-in creative ideas within the ChatGPT Images experience, without requiring users to write a detailed prompt. This approach is designed to lower the barrier to experimentation while still supporting more advanced creative workflows for users who want greater control.

For readers who want to explore additional visual examples and editing workflows, including examples of image editing below, OpenAI has published a gallery of image edits and transformations alongside the announcement.

OpenAI says the new ChatGPT Images model follows instructions more reliably than its initial image-generation release. This improvement supports both more precise edits and more intricate original compositions, helping ensure that relationships between visual elements are preserved as intended.

Text rendering has also advanced, according to OpenAI, with the model better able to handle smaller and denser text. This addresses a long-standing limitation in image-generation systems, where text accuracy and legibility have historically been difficult to maintain.

In addition, OpenAI notes that the model delivers broader quality improvements that make image outputs more immediately usable. These include better rendering of scenes with many small faces and more natural-looking results overall.

Alongside the image model upgrade, OpenAI is introducing a dedicated Images space within ChatGPT, accessible from the sidebar in the mobile app and on chatgpt.com. According to OpenAI, the new hub is designed to make exploring and creating images faster and easier within the ChatGPT experience.

The Images hub includes dozens of preset filters and prompts intended to help users quickly experiment with visual ideas. OpenAI says these presets will be updated regularly to reflect emerging visual trends, supporting everything from small image edits to more complete visual reimaginings.

Taken together, OpenAI says these updates are intended to support a wide range of image creation workflows, from small, targeted edits to more extensive visual reworkings within a single interface, helping users create images that better match their vision.

For developers and businesses, GPT Image 1.5 delivers the same improvements available in ChatGPT Images, with a particular emphasis on image preservation and consistency across edits. OpenAI says the model is stronger at maintaining visual integrity than GPT Image 1, making it better suited for professional and commercial workflows.

According to OpenAI, common use cases include:

OpenAI also says image inputs and outputs in GPT Image 1.5 are priced approximately 20% lower than GPT Image 1, allowing teams to generate and iterate on more images within the same budget.

While OpenAI reports clear improvements across many evaluated examples, the company notes that results are still imperfect and that significant room for improvement remains. The release represents what OpenAI describes as a “meaningful step forward,” with future updates planned to enable finer-grained edits and richer, more detailed outputs across languages.

Q: What did OpenAI release?
A: OpenAI released a new version of ChatGPT Images, powered by a new flagship image generation model, and made it available in the API as GPT Image 1.5.

Q: Who can access it now?
A: OpenAI says the new model is rolling out today to all ChatGPT users and API users globally, while Business and Enterprise access will arrive later.

Q: What’s improved in image editing?
A: OpenAI says edits adhere more reliably to user intent and preserve important details—such as lighting, composition, and people’s appearance—changing only what the user requests.

Q: What types of edits does it support?
A: OpenAI says it supports editing actions including adding, subtracting, combining, blending, and transposing elements.

Q: What’s improved beyond editing?
A: OpenAI says the model is up to 4× faster, follows instructions more reliably, improves text rendering for smaller/denser text, and produces more natural outputs—including better handling of many small faces.

Q: What is the new Images feature inside ChatGPT?
A: OpenAI says it’s a dedicated Images space in ChatGPT (mobile sidebar and chatgpt.com) with preset filters and prompts to make exploring image creation faster and easier.

Q: What is GPT Image 1.5 in the API meant for?
A: OpenAI positions it for the same image creation and editing improvements, with emphasis on consistency for branding and ecommerce workflows like logo preservation and product image catalogs.

Q: Did pricing change in the API?
A: OpenAI says image inputs and outputs are 20% cheaper in GPT Image 1.5 compared with GPT Image 1.

This update matters less because images are faster or cleaner—and more because control and consistency are becoming the default in AI-generated visuals. For marketers, ecommerce teams, and creators, the hardest part of AI image generation hasn’t been creativity—it’s been trust. When a tool changes faces, lighting, or brand elements unexpectedly, it becomes unusable for real work. OpenAI’s emphasis on preserving what users don’t ask to change directly addresses that barrier.

The implications extend beyond professional teams. By reducing the need for prompt expertise and offering preset-driven creative tools, OpenAI is lowering the skill threshold required to produce usable visual assets. That shifts image creation from a specialist task toward a broader, more accessible capability—particularly for small businesses, solo creators, and organizations without in-house design resources.

At a broader level, this signals where AI image tools are headed: not toward replacing human creativity, but toward compressing the distance between idea and execution. As AI systems become more predictable and controllable, the competitive advantage moves away from knowing how to “work the model” and toward knowing what to create and why. For people and teams who rely on visual communication, that shift is likely to matter far more than incremental gains in speed or realism.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.