OpenAI announced the release of GPT-Image-2.6 across ChatGPT, Codex, and its API platform. The model family introduces two API-specific variants: Flare, optimized for low-latency interactive workflows, and Sunburst, designed for detailed image compositions and sustained visual consistency across multiple edit rounds.

The updated models deliver a 50 percent speed increase over previous generations, with Flare reaching render speeds of 6 seconds at low quality settings. The release introduces native SVG vector exporting, animated GIF generation, region-based markup editing, and a new complexity parameter that adjusts planning depth prior to image rendering.

OpenAI reports that overall image generation volume across its platforms has reached over 4 billion images per week, up from 3 billion weekly generations eight weeks prior. The addition of structured formats like SVG targets visual asset workflows for developers and design teams requiring scalable graphics.

Why it matters

  • Native SVG export allows software teams to generate scalable visual assets and UI graphics directly via API endpoints.

  • Variable complexity settings give developers direct programmatic control over generation latency versus layout detail.

  • Faster sub-10-second visual generation capabilities support real-time interactive asset editing applications.

Source: marktechpost.com