Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
A lot has changed in the AI industry in the four months since the release of OpenAI ChatGPT 1.5 images. We’ve had a heated race To build agent toolsthat An unprecedented deal with the Pentagon and The never-ending decline of artificial intelligence.
Now, OpenAI is back in the generative media game. The company announced on Tuesday that it will launch ChatGPT Images 2, a next-generation image model.
ChatGPT Images 2 aims to create text-heavy designs, such as in a matcha ad and a fake magazine cover.
It may seem strange that OpenAI would release a new image model just one month after announcing… The once-popular Sora AI video app has shut down In order to focus on building enterprise-ready “core products.” But it’s clear from how the new model is being built that OpenAI is not backing away from this goal.
ChatGPT Images 2 is designed to produce text-intensive images, including infographics, scientific posters, study guides, and marketing materials. days Sora’s videos are strange and Memes inspired by studio ghibli I finish.
Now, the company is building artificial intelligence that can do what it calls “creative tasks with economic value.”
“The aperture and use cases for visual intelligence are expanding very broadly, and we believe this is critical to ChatGPT’s vision of developing your personal assistant, because your creative assistant is a huge part of who you are as an individual,” Adele Lee, product manager at ChatGPT Images, told reporters at a press conference.
(Disclosure: Ziff Davis, the parent company of CNET, in April 2025 filed a lawsuit against OpenAI, alleging that it infringed Ziff Davis’s copyrights in training and operating its AI systems.)
In these examples, you can see how well ChatGPT Images 2 renders clear text.
OpenAI has been chasing Super dream appa one-stop-shop for all things AI, was created through its Codex platform. ChatGPT Images 2 provides the creative piece of this puzzle.
The new model organically improves typography, iconography and composition to produce more professional AI-powered images. It can generate text in multiple languages. AI image models have struggled to generate clear and correct text. ChatGPT Images 2 is OpenAI’s best model for this yet. Google previously improved text display with Nano Banana Probut even this “best ever” model had difficulty being accurate.
ChatGPT Images 2 is rolling out to all users now. Your creation limit depends on your plan: the more you pay, the more AI images you can create.
Developers using the model in the API can create 2K and 4K images, though these higher resolutions are still in beta and may be wonky. Paying users can also create images using reasoning and reasoning models, which help them search the web for information, compile it into a readable design and double-check their work.
“Image model” doesn’t seem to be quite the correct term for ChatGPT Images 2, even though it is technically correct. ChatGPT doesn’t capture the fantastical surrealism of AI imagery like Midjourney, nor does it offer anywhere near the editing tools of Adobe Firefly.
But it caters to a group of users in the middle of the spectrum from Midjourney’s technical enthusiasts and professional Adobe creators: those who need to create engaging content.
Like Anthropic’s recently released Claude Design, OpenAI’s ChatGPT Images 2 is aimed at working professionals. Teachers can use it to create study guides and illustrated lesson plans. Marketing managers can create social media posts and visual assets.
You can create up to eight images from a single prompt, such as a three-page report, maintaining visual consistency across each.
You can create longer reports using ChatGPT Images 2, and all matching pages.
This is the second half of our AI-generated key lime pie recipe. Notice the visual consistency.
One downside is that if you want to edit an AI-powered photo, you’ll still need to recreate it. With more text-heavy designs, this will likely be necessary, so you’ll be able to implement your credits faster. OpenAI said it focuses on maintaining an iterative and speed-based editing flow to maintain ease of use.
OpenAI’s safety procedures have not changed significantly since the last image model. It still includes metadata through the C2PA standard, so that the origins of AI images can be identified. Offensive and illegal images are still prohibited in OpenAI policies, an important guardrail for AI companies to implement effectively, in light of recent examples of… Deepfakes generated by artificial intelligence and Non-consensual intimate images.
