OpenAI launches ChatGPT Images 2.0 with web‑search and multi‑image generation
Key Points
- OpenAI releases ChatGPT Images 2.0, powered by the new GPT‑Image‑2 model.
- New "thinking" mode lets the generator search the web and analyze uploaded files before creating images.
- Users can generate up to eight coordinated images from a single prompt.
- Resolution increased to 2K; aspect ratios now span 3:1 to 1:3.
- Support for text in Japanese, Korean, Chinese, Hindi and Bengali sees major improvements.
- Thinking capabilities available to ChatGPT Plus, Pro, Business and Enterprise subscribers.
- All ChatGPT users receive core enhancements to realism and style fidelity.
- Launch coincides with heightened competition from Google’s Nano Banana Pro and Microsoft’s MAI‑Image‑2.
OpenAI unveiled ChatGPT Images 2.0, a revamped AI image generator that can browse the web for reference, produce up to eight coordinated images per prompt, and output pictures up to 2K resolution. The upgrade, powered by the new GPT‑Image‑2 model, adds “thinking” capabilities for better detail retention, multilingual text rendering and a broader range of aspect ratios. The features roll out to ChatGPT Plus, Pro, Business and Enterprise users, while all ChatGPT accounts receive core improvements to photo realism and style fidelity.
OpenAI announced today that its AI‑powered image generator, ChatGPT Images, is receiving its most significant upgrade since the tool’s debut last year. Dubbed Images 2.0, the service now leverages a new GPT‑Image‑2 model equipped with what the company calls “thinking capabilities,” allowing the system to pull information from the web before it creates a picture.
When users enable the thinking model, the generator can browse online sources, analyze uploaded files and “reason through the structure of the image before generating.” That extra step lets it preserve specific details users request, produce coherent text within images, and maintain consistent characters, objects and styles across multiple outputs. In practice, a single prompt can yield up to eight images that belong to the same visual narrative—ideal for manga pages, a series of social‑media graphics or room‑by‑room design plans.
Resolution and format options also expand. Images can now be rendered at up to 2,048 pixels on a side and in aspect ratios ranging from ultra‑wide 3:1 to tall 1:3. The upgrade improves the model’s handling of non‑Latin scripts, delivering “significant gains” for text in Japanese, Korean, Chinese, Hindi and Bengali. English and other Latin‑based languages remain strong, but the broader language support marks a notable step forward for creators working in multilingual contexts.
Availability follows a tiered rollout. Subscribers to ChatGPT Plus, Pro, Business and Enterprise plans receive the full suite of thinking capabilities, including web‑search‑enabled image generation. All other ChatGPT users, however, gain access to core enhancements that improve the tool’s ability to capture the defining characteristics of photos, pixel art, cinematic stills and other visual styles. OpenAI says the updates are live for all ChatGPT and Codex users as of today.
The new version arrives amid intensifying competition in the generative‑image arena. Google recently introduced Nano Banana Pro, while Microsoft unveiled MAI‑Image‑2, both of which aim to push speed and fidelity. OpenAI’s last major update in December added faster generation and better photo‑editing features, but the current rollout represents a more expansive leap in functionality.
OpenAI first rolled out ChatGPT Images in 2023, positioning it as a companion to its conversational AI. Since then, the tool has evolved from a novelty to a utility for designers, marketers and hobbyists alike. By integrating web‑search abilities and multi‑image output, OpenAI hopes to make the generator a practical component of creative workflows, not just a one‑off novelty.
Industry observers note that the move underscores a broader trend: AI models are increasingly being given “reasoning” layers that let them reference external data before producing output. Whether that approach will become standard across the sector remains to be seen, but OpenAI’s latest push signals its commitment to staying at the forefront of generative‑image technology.