TL;DR
OpenAI has released a major update to ChatGPT Images powered by GPT Image 1.5. The new model generates images up to 4x faster, follows instructions more reliably, and makes precise edits while preserving original image details like lighting, composition, and facial likeness.
Key Improvements
The update focuses on several capability enhancements:
Precise editing - The model changes only what you ask while keeping other elements consistent. This enables practical photo edits, believable clothing try-ons, and stylistic transformations that retain the essence of original images.
Improved instruction following - The model handles complex compositions more reliably, maintaining relationships between elements as intended.
Better text rendering - Denser and smaller text can now be rendered accurately, demonstrated by examples including complete newspaper layouts with tables and formatting.
Faster generation - Images render up to 4x faster, and users can generate new images while others are in progress.
New Images Experience
Beyond the model upgrade, OpenAI has introduced a dedicated Images section in the ChatGPT sidebar. This includes preset filters and trending prompts to jumpstart inspiration, plus a one-time likeness upload feature for reusing your appearance across future creations.
API Availability
GPT Image 1.5 is available to API developers with 20% lower pricing compared to GPT Image 1. OpenAI highlights particular value for marketing and brand work, including logo preservation across edits and product catalogue generation from single source images.
Early adopters include Wix, who noted the model’s “high-fidelity images with strong prompt adherence” supporting faster concept-to-production workflows.
Looking Forward
For UK businesses using AI for visual content, the update offers meaningful improvements in consistency and speed. The combination of precise editing and faster iteration could accelerate creative workflows—though OpenAI acknowledges the technology still has “significant room for improvement.”