Introducing ChatGPT Images 2.5

OpenAI unveiled ChatGPT Images 2.5, a new image model designed to deliver sharper details, faster generation, and more reliable editing compared to its predecessor. The update improves natural lighting, texture richness, and subject preservation in reference photos, while reducing generation latency by up to 50%. Users across ChatGPT, ChatGPT Work, and Codex platforms can now access these improvements on desktop, mobile, and web interfaces.

The model introduces new features to streamline the creative process, including Sketch, which allows users to draw directly in ChatGPT as a visual reference for final images. Templates for popular formats like flyers and product photos simplify starting points, while comment placement on images enables more focused editing. Users can also share prompts alongside images to let others recreate variations with their own details.

For developers, OpenAI introduced two new API models: GPT-Image-2.5 Flare, which mirrors the quality and speed improvements of Images 2.5, and GPT-Image-2.5 Sunburst, offering higher precision for detailed creative work at the cost of longer generation times. Both models aim to enhance reference-led workflows, ensuring variations remain anchored to original sources.

Images 2.5 excels at following complex editing instructions across multiple turns, maintaining consistency in lighting, textures, and subject features even after repeated edits. The model better handles real-world visual information, transparent backgrounds, and artistic styles, making it more dependable for tasks like generating on-brand assets or structured UI concepts. Developers and businesses can rely on it to preserve visual direction and composition in large-scale creative projects.