OpenAI launched a example of a new generation of AI on Tuesday, named ChatGPT Images 2.0. This format can generate more than one image from a single prompt, such as an entire learning booklet, as well as output text, including in non-English languages such as Chinese and Hindi. This version is available worldwide ChatGPT and Codex users, and a more powerful version is available to paying customers.
When any major AI company releases a new photo design, it can revive interest and increase usage, especially if social media users follow a memorable trend, changing their own photos. Last year, Google’s launch of the Nano Banana model was an important moment for the company, especially when users started to print. hyperrealistic sculptures themselves online. Earlier this year, ChatGPT Images made waves on social media as users shared them AI based animations.
What’s the Difference?
Since the new model can tap into the “reasoning” ability of ChatGPT, Photo 2.0 can search for recent news on the Internet and provide more than one photo at a time. In fact, the bot can use additional steps to produce more detailed generations from a single prompt. Image 2.0 also has a more recent cut-off date: December 2025.
This also means that the output from the new model is more granular. For example, I created an infographic with the weather forecast for San Francisco for the next day, along with the activities that should be done. The image produced by ChatGPT included accurate weather information for a rainy day, as well as accurate depictions of the Ferry Building, Castro Theater, the houses of the Painted Ladies, and the Transamerica Pyramid.
Additionally, Photo 2.0 can be further customized for users who want a unique balance of photo results. The new model can produce images ranging from 3:1 in width to 1:3 in height, and users can adjust the image size as part of their inspiration with the AI tool.
First Impressions
After a few hours of creating images and new designs, I was generally impressed with the presentation capabilities of the text, in English at least. Not so long ago, textual image output, from any of the major formats, often included many misspelled characters or words with additional misspelled characters. ChatGPT struggled to label image exactly two years ago, so the cleaner, more complex results from Image 2.0 are a sign of continued improvement. Google has also focused on improving image results with text in them recent iterations of Nano Banana.
