OpenAI released GPT-Image-1.5 on Tuesday, an upgraded image generator that the company claims runs four times faster than its predecessor and handles edits without destroying what you've already built. The model is available now to all ChatGPT users and through the API.
The pitch: ask for specific changes to a generated image, like adjusting lighting or swapping an element, and the model actually follows through. Previous versions had a habit of reinterpreting the entire image when users requested small tweaks. OpenAI says GPT-Image-1.5 preserves facial likeness, lighting, and composition across sequential edits. Whether it delivers consistently remains to be tested at scale.
This wasn't the original timeline. OpenAI had reportedly planned to ship the new image model in early January but accelerated the release. The timing tracks with CEO Sam Altman's leaked "code red" memo from last month, which outlined plans to counter Google after Gemini 3 and Nano Banana Pro took the top spots on LMArena. Google's chatbot hit 650 million monthly users in October, up from 450 million in July.
Along with the model upgrade, ChatGPT gets a dedicated "Images" section in the sidebar with preset filters and trending prompts. Fidji Simo, OpenAI's CEO of applications, described it as working "more like a creative studio." API users also get a price cut: 20% cheaper for image inputs and outputs.
The Bottom Line: OpenAI shipped its first major image model update since April, eight months after GPT-Image-1, with Google's image tools clearly in the crosshairs.
QUICK FACTS
- GPT-Image-1.5 available to all ChatGPT users and API (December 16, 2025)
- Speed: 4x faster image generation (company-reported)
- API pricing reduced 20% for image inputs/outputs
- Previous model (GPT-Image-1) launched April 2025
- Google Gemini: 650 million monthly users as of October 2025
