OpenAI Launches ChatGPT Images 2.5 with Direct Sketching and 50% Speed Upgrade

Artificial intelligence research lab OpenAI has released ChatGPT Images 2.5, representing a significant upgrade to its native visual creation engine across ChatGPT, ChatGPT Work, and Codex platforms. Five months after debuting version 2.0, the latest release delivers major advancements in generation speed, visual fidelity, and precise canvas editing. OpenAI stated that image generation latency has been reduced by up to 50 percent compared to the previous model, enabling creators to render visual concepts and iterate on prompt designs much faster. Among the flagship capabilities in Images 2.5 is a brand-new feature called Sketch, accessible by invoking '@Sketch' within chat conversations. This allows users to draw quick rough outlines or hand-sketched layouts directly on their screens to serve as direct structural reference for the final high-definition AI image. Additionally, the new model renders more realistic lighting, richer surface textures, better subject retention across multi-turn prompts, and improved handling of transparent backgrounds and non-Latin character text. Along with the consumer rollout on web and mobile apps, OpenAI has introduced two new API models—GPT-Image-2.5 Flare and GPT-Image-2.5 Sunburst—designed for developers building high-volume image applications and rapid visual search tools. Users can now also place precise comment markers directly on generated images to guide multi-step edits without altering the rest of the canvas composition. This release underscores OpenAI's ongoing push to convert AI chatbots into comprehensive, multi-modal creative studios.