OpenAI has announced the official launch of its next generation image generation technology. The new GPT-Image-2.6 system is now accessible simultaneously via the company API, the ChatGPT platform and the Codex programming assistant. This release indicates an effort to connect generative visual capabilities directly into automated workflows and developer environments. Software engineers and digital artists can now tap into the updated generation pipeline across existing development setups without relying on external bridge tools.
To accommodate different practical demands, the model is distributed in two specialized variants. The edition designated as Flare concentrates on delivering high detail density and refined visual fidelity for demanding compositional tasks. In contrast, the model titled Sunburst is explicitly engineered for high speed and lower processing latency. This split allows teams to select the appropriate engine depending on whether their current use case prioritizes instant user feedback or fine-grained rendering quality.
A central technical addition in this release is the native generation of looping GIF animations through a single API query. Previous workflows typically required multi-step generation pipelines, frame extraction and external stitching software to achieve smooth animated loops. By returning playable, cyclical short-form files from one request, the update dramatically reduces the engineering overhead needed to build dynamic assets for messaging services, user interfaces and social media.
Equally relevant for digital designers is the introduction of direct vector exports formatted as SVG files. Rather than limiting outputs to static raster images, the model can now generate scalable vector data that can be enlarged without quality degradation. This output format integrates smoothly into standard graphic design software, bridging the gap between artistic image prompts and technical interface requirements. UI designers can now extract icons, logos and schematic illustrations that fit directly into responsive web layouts.
The update also improves in-place visual revisions through a mechanism designated as Circled Edits. This markup-assisted editing framework lets users draw physical circular marks on specific segments of an image while passing text instructions for localized changes. By grounding prompts in clear visual markers, the model minimizes ambiguous edits and protects the rest of the canvas from unwanted drift. Together with SVG exports and animation features, these changes push generative models further toward versatile production tools.