Key Highlights:
- GPT Image 1.5 generates higher-fidelity images with better prompt accuracy while keeping lighting, composition, and fine details consistent.
- ChatGPT now supports more precise image editing, improved text rendering, and realistic visual transformations across multiple edits.
- A new Images section with preset filters and prompts is rolling out globally for all ChatGPT and API users.
After announcing code-red companywide, OpenAI has started pushing its latest products to take on its rivals. And, today, the company has dropped GPT Image 1.5, its new AI image model that it claims “generates high-fidelity images with strong prompt adherence, preserving composition, lighting, and fine-grained detail.”
The new image model is better at preserving details within the images
A new version of ChatGPT Images is rolling out today, powered by the company’s latest flagship image model, and it promises faster, more precise image creation and editing. Whether you’re starting from scratch or making edits to an uploaded photo, the latest image model follows instructions more reliably, preserving the details that matter.
When I say details, here I mean lighting, composition, and how people look in the generated image. Per the company, details remain consistent, even with multiple edits. This means ChatGPT can now generate more realistic clothing try-ons, improved stylistic filters, and conceptual transformations that still feel true to the original.
The new model also processes text rendering and small details significantly better. The editing tools have also been updated to help you add, remove, combine, or transpose elements with precision. In short, this update makes ChatGPT a pocket-sized photo studio of some kind.
You may also like: OpenAI’s new GPT-5.2 model takes on Google’s Gemini 3
A dedicated Images feature with suggestions
Besides the GPT Image 1.5 model, OpenAI has also introduced a dedicated Images feature in ChatGPT. You can access it via the sidebar on the mobile app and chatgpt.com. It comes with dozens of preset filters and prompts for inspiration.
The latest update is available to all ChatGPT users and API users worldwide, and it works seamlessly across models. Have you tried the new image model yet? If yes, what do you think about the generated images? We’d love to hear; drop your thoughts in the comments below.
