In March, OpenAI released a major upgrade to ChatGPT’s image generation capabilities, powered by a new model that became a viral hit almost overnight. This single update brought millions of new users to ChatGPT, with users creating more than 700 million images in just one week. In April, OpenAI extended the same image generation technology to developers with the launch of the gpt-image-1 API.
In September, Google introduced a new image generation model codenamed Nano Banana, which quickly went viral and significantly boosted Gemini’s reach. The Nano Banana model outperformed gpt-image-1 in both image creation and editing, and it gained strong adoption among developers worldwide. Last month, Google followed up with Gemini 3 Pro Image, codenamed Nano Banana Pro, which leverages Gemini’s state-of-the-art reasoning and real-world knowledge to produce more accurate images than ever before.
Today, OpenAI has released gpt-image-1.5, a new image generation model aimed at competing more directly with Google’s Gemini Nano Banana lineup.
As expected, the gpt-image-1.5 model offers more precise image editing and better preservation of logos and faces. It is also significantly stronger at instruction following and adherence to user prompts. OpenAI has further improved text rendering in generated images, particularly for denser and smaller text.
For developers, gpt-image-1.5 is 20% cheaper for image inputs and outputs compared to gpt-image-1. As before, developers can further control image generation costs using the quality parameter. OpenAI claims the new model delivers strong performance even when the quality setting is set to low.
For general consumers, the new model now powers the ChatGPT image generation experience. OpenAI highlighted that it can make accurate edits while keeping details such as lighting, composition, and people’s appearance intact, delivering up to 4× faster performance compared to the previous model.
OpenAI is also introducing a new dedicated Images hub in ChatGPT, available in the mobile app and on chatgpt.com. Users can quickly get started with image generation using a range of preset filters and prompts. OpenAI noted that these presets and prompts will be updated regularly to reflect emerging trends.
The new image generation model and the ChatGPT Images experience are rolling out to all ChatGPT users.