GPT Image 1.5
2025-12-16 · Publisher : OpenAI · Domain : Artificial Intelligence
OpenAI launches a new ChatGPT Images powered by GPT Image 1.5, its flagship image model, more precise in edits, more faithful to instructions and up to four times faster, also available in the API.
What changes
- More precise edits: the model changes only what is requested, preserving lighting, composition and people's appearance between versions.
- Better instruction following, text rendering and creative transformations, with preset styles in a new Images space in ChatGPT.
- Generation up to 4 times faster.
- More stable preservation of logos and brand visuals, useful for marketing and e-commerce catalogs.
- Rollout to all ChatGPT users; Business and Enterprise for the Images space afterwards.
### Specifications - Text and image input, image (and text) output. - API: gpt-image-1.5 (snapshot 2025-12-16) and chatgpt-image-latest; added to the image generation tool of the Responses API on December 19, 2025. ### Pricing Image inputs and outputs 20% cheaper than GPT Image 1. Price per million tokens (USD), according to the documentation. | | Input | Cached input | Output | | --- | --- | --- | --- | | Text | 5 | 1.25 | 10 | | Image | 8 | 2 | 32 | Per square image: about 0.009 (low), 0.034 (medium), 0.133 (high) dollars. ### Benchmarks No numerical score published by the vendor in the blog post. ### Key takeaways GPT Image 1.5 replaces gpt-image-1 for image editing in production. Consistency across successive edits is the main contribution for creative teams.
Links
Tags : Image, OpenAI, GPT Image, ChatGPT, Multimodal