OpenAI Releases ChatGPT Images 2.5 with Sunburst and Flare API Models
OpenAI has officially released ChatGPT Images 2.5, featuring up to 50% lower image generation latency, improved subject preservation, and interactive features like sketch input and in-image commenting. Alongside the ChatGPT update, OpenAI introduced two API models: `gpt-image-2.5-sunburst` for high-precision visual workflows and `gpt-image-2.5-flare` for fast everyday generation. This update directly addresses major pain points in generative image models, including detail degradation over iterative edits and rendering speed. By offering separate fast and high-precision API models, developers can better tailor image generation to specific latency or fidelity requirements. The new in-image commenting capability lets users pin targeted prompt edits to specific regions without altering surrounding details. For API integrations, `gpt-image-2.5-flare` serves as the lower-latency default, while `gpt-image-2.5-sunburst` trades longer generation times for finer control and higher detail accuracy.
## BACKGROUND
Multi-turn image editing refers to the process of iteratively refining an AI-generated visual across continuous prompts in a single conversation. Previous text-to-image models often struggled to maintain subject consistency, lighting, and style when applying multiple sequential modifications.