Precise, Faster, Cheaper: OpenAI Unleashes GPT-Image-1.5 to Reclaim the Visual AI Crown

OpenAI has released GPT-Image-1.5, claiming improvements that address speed, controllability and cost in the visual AI space, in response to Google Nano Banana’s resurgence. The article notes that GPT-Image-1.5 is four times faster than its predecessor and aims to deliver real-time feedback while dramatically cutting trial-and-error costs, with image input and output costs via the API reduced by 20% to $8 per million input tokens.

OpenAI has also introduced a new image-creation entry point in ChatGPT’s web and mobile apps and bundled a lightweight, photo-editing style interface featuring filter libraries and prompt templates, designed to make prompts easier for everyday users. Demonstrations reportedly show improved character consistency and the model’s ability to change only selected aspects of an image while leaving other regions untouched, addressing the familiar “butterfly effect” in AI image editing.

The release is framed as a pragmatic, commercially focused move, with the text noting OpenAI’s pursuit of tighter workflows to attract enterprise customers such as Wix and Canva. Published on 17 December 2025, the piece situates the launch amid ongoing competition in the visual generation space.