Latest models
- GPT Images 2.5 (ChatGPT Images 2.5) (September 8, 2026) Current
OpenAI's latest image update in ChatGPT and the API, described as faster and sharper, with better fidelity and tools for creating images. Source: OpenAI announcement ↗ - GPT Image 2 (April 21, 2026) Current
OpenAI's image model in the API and Codex, with stronger editing, better layouts and text rendering, and output up to 2K. Source: OpenAI developer docs ↗
Checked October 7, 2026. See the full AI model tracker.
What we like
- OpenAI says it is stronger at editing, layouts and text rendering
- Available in ChatGPT, so there is nothing separate to learn
- Supports resolutions up to 2K in the API
Watch out for
- Access and limits depend on your ChatGPT plan or API use
- Defaults can feel less stylised than Midjourney
- OpenAI's claims are its own; test with your own prompts
GPT Image is the name for OpenAI’s current image models. They arrived in ChatGPT and the API after DALL-E, which OpenAI removed from the API on May 12, 2026 (see DALL-E).
What is current
- GPT Image 2 (
gpt-image-2) launched on April 21, 2026 in the API and Codex. OpenAI describes stronger editing, better layouts, improved text rendering and more reliable instruction-following, with resolutions up to 2K. - GPT Images 2.5 (ChatGPT Images 2.5) followed on September 8, 2026 in the API and ChatGPT, which OpenAI describes as faster and sharper with improved fidelity.
See the model tracker for sources.
How prompting works
Write full sentences. Say what the subject is, where it is, the light and the style, and put any text you want in quotation marks. For edits, upload an image and say what to keep and what to change.
Who should use it
People who already use ChatGPT and want reliable results from detailed instructions, including layouts and text. Compare it in Midjourney vs GPT Image and Ideogram vs GPT Image.
Contribute your creativity