OpenAI's image generator can now search the web to create better pictures — available to ChatGPT paid users

OpenAI released ChatGPT Images 2.0, powered by a new GPT Image 2 model, that can search the internet while generating images. The feature is rolling out now to ChatGPT Plus, Pro, Business, and Enterprise subscribers.

When you enable the 'thinking' mode (a feature where the AI spends extra processing time reasoning through your request), the image generator can look up current information online — meaning it can create images based on recent events, specific people, or detailed references that wouldn't be in its training data. It also follows written instructions more precisely and can now generate readable text within images.

For content creators, marketers, and professionals who use AI images for work, this eliminates the friction of having to manually describe or reference current events — you can now ask for an image of 'this week's trending TikTok aesthetic' or 'a product announcement in the style of Apple's recent keynotes' and it will pull real context from the web.

The feature requires a paid ChatGPT subscription (Plus starts at $20/month, or higher tiers for teams and organizations). It's available immediately for subscribers to try.