OpenAI launches Images 2.6 with GIFs and image reference
OpenAI has released Images 2.6, an update to its image generation platform that expands the output format options and refines the overall toolset. The most notable additions are animated GIF support and image reference input, the latter allowing users to supply an existing image to guide the style or composition of a new generation. These two features together move the platform closer to a more flexible, multi-modal creative workflow.
On the API side, the update introduces two new models - Flare and Sunburst. While specific technical details on how these models differ architecturally remain limited, their introduction suggests OpenAI is segmenting its image generation capabilities to give developers more targeted options depending on the use case. This kind of model tiering is common in the generative image space and allows for trade-offs between speed, quality, and cost.
Faster image generation is also part of the 2.6 release, which is a practical improvement for applications that rely on near-real-time output, such as interactive tools or high-volume pipelines. Alongside speed gains, the update includes more precise editing controls, which likely benefits workflows that require iterative refinement - adjusting specific regions or elements of an image without regenerating the whole composition.
Taken together, the Images 2.6 update reflects a steady expansion of OpenAI's image platform rather than a sharp departure from what came before. The addition of GIF output is particularly worth noting, as animated formats have been a gap in many text-to-image tools. Whether the GIF support extends to full prompt-driven animation or is limited to shorter, looping outputs remains a detail worth watching as more users test the release.
