OpenAI announced the release of ChatGPT Images 2.5, describing it as a new state-of-the-art image model that delivers sharper details, faster generation, and more precise editing capabilities. According to the announcement, the model produces more natural lighting and richer textures, preserves subjects from reference photos more reliably, and follows editing instructions more consistently across multiple turns. The company also said that image generation latency has been reduced by up to 50 percent compared with the previous Images 2.0 model, allowing users to create and refine concepts more quickly. In the ChatGPT interface, several new product features accompany the model update. A Sketch function lets users draw directly in the chat as a visual reference for the final image; typing “@Sketch” activates the tool. Templates for common formats such as flyers, posters, and product photos provide a starting point, while the ability to place comments on images and share the prompts used to generate them aims to streamline collaborative workflows. These features are available to all ChatGPT, ChatGPT Work, and Codex users across desktop, mobile, and web platforms. For developers using the Images API, OpenAI introduced two new model variants. GPT‑Image‑2.5 Flare is positioned as the default choice for most applications, offering the same quality, editing, and speed improvements as the consumer‑facing model and achieving up to 50 percent lower latency than the earlier GPT‑Image‑2. GPT‑Image‑2.5 Sunburst targets premium visual workflows that require tighter control across edits, such as production‑ready campaign creative or polished product imagery, and provides an extra level of precision at the cost of longer generation times. OpenAI emphasized that Images 2.5 improves reference‑based workflows, making it easier to transform familiar subjects into new settings, visual styles, or compositions while keeping distinctive features intact. The model is said to handle complex visual instructions, transparent backgrounds, and diverse artistic styles more reliably, reducing drift over multiple editing turns—a benefit for teams that need to iterate on brand‑consistent assets or UI concepts. Safety measures remain in place, with prompt and image checks, C2PA metadata, and invisible watermarking to help identify AI‑generated content. A system card detailing evaluations and safeguards is linked in the announcement. The rollout of ChatGPT Images 2.5 began today for all eligible users, and pricing details for the new API models are available via the provided link.
