OpenAI ships ChatGPT Images 2.6 with GIFs, reusable references and two faster API models
OpenAI has rolled out ChatGPT Images 2.6 to all ChatGPT and Codex users, adding native animated-GIF creation and saved reference images for people, pets and characters. Two new API models, GPT-Image-2.6 Flare for speed and Sunburst for quality, promise 50 per cent faster generation as weekly image volume passes four billion.
OpenAI has upgraded its image stack with ChatGPT Images 2.6, available from 4 November to every ChatGPT and Codex tier on web, desktop and mobile. The consumer-facing additions are animated GIF generation from a prompt or an existing image, with control over speed and the ability to edit individual frames, and reusable reference images that keep a saved likeness of a person, pet or character available without re-uploading it each time.
For developers, two new models join the API. GPT-Image-2.6 Flare is tuned for speed and, according to OpenAI, can return an image in as little as six seconds at lower quality settings. GPT-Image-2.6 Sunburst is the quality-focused sibling, offering higher-fidelity generation and stronger consistency across successive edits, along with image markup and SVG export. OpenAI's headline claim is generation that is 50 per cent faster than GPT-Image-2.5 overall.
The company did not disclose pricing for Flare and Sunburst in the announcement, and reports that the API rates match Images 2.5 remain unconfirmed. What OpenAI did disclose is scale: users now generate more than four billion images a week across ChatGPT and the API, up from the three billion figure it cited in September.
The release is incremental rather than a leap, but the direction is telling. Consistency across edits, SVG output and persistent references are the features that move image generation from novelty towards production use in design, marketing and documentation pipelines, where the same character or product must appear identically across dozens of assets.
Why it matters
Image models are quietly becoming infrastructure. Four billion images a week is a volume that makes speed and per-image cost, not raw quality, the decisive competitive variables, and a six-second fast path with a quality tier alongside it is how that market will be segmented.
Indicops view
For Indian public-sector communications teams and enterprises producing multilingual collateral, consistent reference-based generation and SVG export are genuinely useful, provided brand and likeness governance is in place. We would pair any adoption with clear rules on whose likeness may be saved, watermarking of generated assets and a review step before publication, since the same features that make campaigns efficient make misuse easier.
More in Models
Back to the newsroomGoogle launches Gemini 4 Hydrogen, a $1.50-per-million-token model for coding and agents
Google has released Gemini 4 Hydrogen, a faster and cheaper tier of its Gemini 4 family aimed at everyday coding and agentic work. It costs $1.50 per million input tokens and $7.50 per million output tokens across a 1M-token window, and is rolling out through the Gemini API, AI Studio and consumer products.
Alibaba opens a two-week preview of Qwen4-Max with a 1M-token context, but no benchmarks or weights
Alibaba Cloud has opened a time-limited preview of Qwen4-Max, its next flagship model, through the Token Plan only. It offers a 1M-token context and up to 393,216 output tokens, but the documentation lists no pricing, benchmarks or downloadable weights, and the model retires on 25 November.
Anthropic releases Claude Fable 5.2 and Mythos 5.2, 30 per cent faster and 20 per cent cheaper
Anthropic has released Claude Fable 5.2 for general use and Mythos 5.2 for vetted cyber and life-sciences organisations. The company says output is up to 30 per cent faster than the 5.1 generation, at $8 per million input tokens and $40 per million output tokens, with lower rates of safeguard blocks.