ChatGPT Images gets upgrade as OpenAI touts faster generation and accuracy

Generating images through ChatGPT is set to get easier, faster, and more accurate as OpenAI announced the release of ChatGPT Images⁠ powered by its new flagship image generation model, as well as a new Images feature⁠ within ChatGPT.

OpenAI claimed that the model showed improvements across many use cases but confirmed there were still imperfections in images that needed improvement.

However, the ChatGPT-maker touted better consistency across lighting, composition, and people’s appearances when making AI edits to uploaded images.

“This unlocks results that match your intent—more useful photo edits, more believable clothing and hairstyle try-ons, alongside stylistic filters and conceptual transformations that retain the essence of the original image,” said OpenAI in its blog post.

For users who are generating images with text prompts, OpenAI highlighted that the new model would follow instructions more reliably, create increasingly natural-looking facial outputs, and render text better than before to deliver denser and smaller letters.

There will also be a dedicated home space for Images⁠ in ChatGPT.

Users of past AI image generation models have criticised unnatural details in human bodies, visibly AI-generated typefaces, and a yellow hue to images that make them hard to deploy for creative or commercial use.

OpenAI announced that the new Images model is rolling out in ChatGPT for all users and that it can be accessed in the API as GPT Image 1.5. Business and Enterprise access are set to come later.

“We believe we’re still at the beginning of what image generation can enable. Today’s update is a meaningful step forward with more to come, from finer-grained edits to richer, more detailed outputs across languages,” said OpenAI in its blog post.