OpenAI introduced Images 2.6 on November 4, 2026 in ChatGPT, Codex, and the API. The release adds native animated GIF generation, claims image generation that is 50% faster than before, and brings two API models: GPT-Image-2.6 Flare for speed and GPT-Image-2.6 Sunburst for higher detail and edit consistency.
Key Facts
| Item | Detail |
|---|---|
| Product | Images 2.6 |
| Chat surfaces | ChatGPT and Codex, all subscription tiers, desktop, mobile, and web (global, same day) |
| New ChatGPT capability | Animated GIFs from text prompts or uploaded photos; users can change playback speed and edit frames (per The Verge, citing OpenAI) |
| Speed claim | About 50% faster image generation than prior Images; Flare at low quality can finish in as little as six seconds |
| API models | gpt-image-2.6-flare (fast iteration) and gpt-image-2.6-sunburst (detail / edit consistency at medium or high quality) |
| API extras | Image markup for targeted edits, SVG export, complexity parameter, GIF output on generations |
| GIF API knobs (docs) | output_format: "gif"; frame_count 2–64 (default 9); frame_duration_ms 20–2000 (default 200); image edits endpoint does not support GIF output |
| Volume claim | More than 4 billion images per week across ChatGPT, Codex, and the API |
| Safety / provenance | System card published the same day; GIF-specific moderation; C2PA metadata and SynthID watermarking |
Background
Before Images 2.6, ChatGPT could not generate a GIF in a single native step, The Verge reported. OpenAI's announcement shows GIF workflows for brand mascots, social reactions, and storyboards, and says image markup already available in ChatGPT now reaches the API so developers can mark where an edit should land. A complexity control sets how much planning the model does before rendering; quality still controls visual detail.
The ChatGPT Images 2.6 System Card says Flare is optimized for faster generation, confirms GIF generation and photo markup in ChatGPT and the API, and SVG output in the API. It also describes GIF-specific safety checks that score a frame grid plus caption, because motion can carry meanings that a single still frame does not.
Reactions and Impact
Early-access quotes on OpenAI's post include OpenArt CEO Coco Mao on GIF as social media's "native language," and Figma Weave Creative Director Itay Schiff on adding the model to Weave workflows. The Verge and The Decoder independently confirmed the ChatGPT GIF controls and the API Flare / Sunburst split. OpenAI pointed readers to API pricing pages rather than listing dollar rates in the announcement.
What We Don't Know Yet
- Per-image dollar prices for Flare and Sunburst (announcement links pricing docs; we did not restate unverified figures)
- How DesignArena or other third-party image leaderboards rank Images 2.6 (not verified on primary sources for this story)
- How often GIF-specific moderation misses motion-based policy violations (system card reports evaluation recall on a fixed set, not production rates)