OpenAI has released a new version of ChatGPT Images, powered by its flagship GPT-Image-1.5 model, promising faster image generation, more accurate edits, and stronger instruction following.
The update, announced on December 16, 2025, is rolling out globally across ChatGPT and the API, marking a major step forward in AI-powered image creation and editing.
The release introduces a redesigned Images experience inside ChatGPT, enabling users to generate and edit images more intuitively, while preserving critical details such as facial likeness, lighting, and composition.
OpenAI says the new model generates images up to four times faster, which allows users to iterate and explore creative ideas with significantly less delay.
Introducing ChatGPT Images, powered by our flagship new image generation model.
– Stronger instruction following
– Precise editing
– Detail preservation
– 4x faster than beforeRolling out today in ChatGPT for all users, and in the API as GPT Image 1.5. pic.twitter.com/NLNIPEYJnr
— OpenAI (@OpenAI) December 16, 2025
What Is New in the Latest ChatGPT Images Release?
The centerpiece of the update is GPT-Image-1.5, OpenAI’s most capable general-purpose text-to-image model to date.
According to OpenAI, the model delivers higher fidelity results, improved text rendering, and more reliable instruction adherence than its predecessor.
The update applies whether users are creating images from scratch or editing existing photos.
OpenAI emphasized that the model now changes only what users ask for, while keeping surrounding details intact across multiple edits, addressing a long-standing challenge in AI image editing.
Alongside the model upgrade, OpenAI has launched a dedicated Images space within ChatGPT, designed to make image exploration faster and more accessible through preset styles, trending prompts and visual inspiration tools.
How Does GPT-Image-1.5 Improve Image Editing Accuracy?
One of the most significant improvements highlighted in the announcement is precision editing.
When users upload an image and request changes, GPT-Image-1.5 adheres more closely to intent, modifying specific elements without altering unrelated parts of the image.
What types of edits does the model handle better? The model excels at multiple editing operations, including:
- Adding or removing objects
- Combining or blending visual elements
- Transposing styles while preserving composition
- Maintaining facial likeness and visual consistency across edits
These capabilities enable more believable clothing and hairstyle try-ons, cleaner background changes and stylistic transformations that retain the essence of the original image.
OpenAI describes this as turning ChatGPT into a “creative studio in your pocket,” capable of both practical image fixes and expressive reimaginings.
How Do Creative Transformations Work in the New Images Model?
Beyond practical edits, GPT-Image-1.5 shows noticeable gains in creative transformations.
The model can change visual styles, layouts, and even text placement while preserving key details.
OpenAI says these transformations work across both simple and complex concepts, allowing users to experiment with movie posters, illustrations, advertisements, and conceptual art.
Importantly, many of these transformations can now be triggered without writing a detailed prompt, using preset ideas inside the new Images experience.
This shift lowers the barrier to creative exploration, especially for non-designers who want high-quality visual output without mastering prompt engineering.
Has Instruction Following Improved Compared to Earlier Models?
Instruction following is a core area where GPT-Image-1.5 outperforms earlier versions.
OpenAI demonstrated this with structured prompts that require precise relationships between multiple elements, such as creating multi-row image grids with specific objects in each cell.
The new model maintains layout integrity, object placement and textual accuracy more reliably than GPT Image 1.0.
This improvement enables more complex compositions where spatial relationships and instructions must be followed exactly, a requirement for professional design and educational use cases.
How Does the Model Handle Text Rendering Inside Images?
Text rendering has been another persistent weakness in image generation models. GPT-Image-1.5 takes a measurable step forward in this area.
The model is capable of rendering denser and smaller text with improved legibility, including markdown-style layouts, infographics, and newspaper-like designs.
OpenAI showcased examples where entire articles, tables, and formatted text blocks were rendered inside images while preserving content structure and accuracy.
While OpenAI acknowledge that text rendering is still imperfect, the improvement makes the model more viable for real-world design tasks such as posters, menus, ads and educational visuals.
What Is the New Images Creation Space in ChatGPT?
To complement the model upgrade, OpenAI introduced a dedicated Images experience inside ChatGPT, accessible through the sidebar on chatgpt.com and the mobile app.
This space allows users to explore image ideas more quickly, using:
- Preset filters and styles
- Regularly updated trending prompts
- Faster parallel image generation
A notable addition is a one-time likeness upload, allowing users to reuse their appearance across future creations without repeatedly uploading photos.
OpenAI says this streamlines workflows for creators experimenting with personal branding or consistent visual identities.
What Additional Quality Improvements Does GPT-Image-1.5 Deliver?
OpenAI reran many examples from its earlier image generation launch to benchmark progress.
The results showed improvements in rendering complex scenes, handling many small faces, and producing more natural-looking outputs.
At the same time, OpenAI acknowledged the remaining limitations. Certain styles, multilingual text rendering and scenes involving multiple faces still show inconsistencies.
The company described the release as meaningful progress rather than a finished solution, noting that further improvements are planned.
How Is GPT-Image-1.5 Available to Developers and Businesses?
GPT-Image-1.5 is available immediately through the OpenAI API, delivering the same improvements as ChatGPT Images.
OpenAI highlighted that the model offers stronger preservation of branded logos and key visuals across edits, which makes it suitable for marketing, ecommerce, and design automation.
Image inputs and outputs are now 20% cheaper compared to GPT Image 1, allowing developers to generate more images within the same budget.
OpenAI says enterprises and startups across creative tools, ecommerce, and marketing software are already adopting the model.
What Early Feedback Are Enterprise Users Sharing?
Companies including Wix, Canva, Figma Weave, Envato, and others are already integrating GPT-Image-1.5 into their workflows.
Wix described the model as delivering high-fidelity images with strong prompt adherence, preserving composition and lighting while supporting faster concept-to-production cycles.
This early feedback positions GPT-Image-1.5 as a competitive flagship model for professional image generation, particularly where consistency and reliability matter.
Key Takeaways
- OpenAI released GPT-Image-1.5, powering the new ChatGPT Images experience worldwide.
- Image generation is up to 4× faster, enabling rapid creative iteration.
- Edits now preserve likeness and composition, changing only what users request.
- A new Images workspace simplifies creation with presets, trends, and reusable likenesses.
- Text rendering and instruction accuracy improved, supporting complex layouts and visuals.
- The API is cheaper and more reliable, with 20% lower image costs and better brand consistency.
Dipti Arora
AuthorDipti Arora is a Senior Content Writer with over seven years of experience creating impactful content across Digital Marketing, SEO, technology, and business domains. She has a strong background in managing news verticals and delivering editorial excellence. Dipti has contributed to leading publications such as The Times of India and CEO News, where her research-driven storytelling and ability to simplify complex subjects have consistently stood out. She is passionate about crafting content that informs, engages, and drives meaningful results.