What Is ChatGPT Images 2.6? GIFs, SVG, 50% Faster

Meet ChatGPT Images 2.6, the GIF-and-SVG Image Upgrade Running 50% Faster

ChatGPT Images 2.6 is an image generation and editing update developed by OpenAI that adds animated GIF output, saved reference images, markup-guided edits and SVG vector export, while generating images 50% faster than the previous version. OpenAI released it on November 4, 2026 in ChatGPT, Codex and the API. Developers access it through two API models: GPT-Image-2.6 Flare, which can return an image in as little as six seconds at low quality, and GPT-Image-2.6 Sunburst, which targets higher detail and consistency across edits. It is available on every ChatGPT and Codex tier across web, desktop and mobile.

Key Takeaways

• ChatGPT Images 2.6 launched on November 4, 2026, at 3:35 PM PT, in ChatGPT, Codex and the API.
• OpenAI states the update generates images 50% faster than before.
• GIF Mode turns a prompt or a still image into an animated loop with adjustable speed and per-frame editing.
• Saved References store images of yourself, pets or characters for reuse without re-uploading.
• The API splits into two models, Flare (speed) and Sunburst (quality and edit consistency), reportedly at the same $30 per 1M image output tokens.
• OpenAI says users now create 4 billion images per week across ChatGPT, Codex and the API, more than half a billion per day.

What Is ChatGPT Images 2.6?

ChatGPT Images 2.6 is the November 2026 version of OpenAI's image system, and it sits in a fast release cadence. Images 2.0 arrived in April 2026 with a model that reasons before drawing, and Images 2.5 introduced the Flare and Sunburst split in September. Version 2.6 keeps that architecture and expands what the model can output and how precisely you can steer edits.

The name covers two surfaces. In ChatGPT and Codex, "Images 2.6" is the image experience you use in conversation: describe an image, upload a reference, or keep editing across turns. In the API, developers pick one of two named models when they send a request. Third-party documentation reviews list the model IDs as gpt-image-2.6-flare and gpt-image-2.6-sunburst, with no dated snapshots at launch.

OpenAI's announcement frames the release around faster iteration and keeping characters, products and brand details consistent as you edit. The status is released, not preview, though GIF creation is reportedly rolling out gradually.

ChatGPT Images 2.6 is best understood as an output-format upgrade: the same reasoning image model, now producing motion (GIF) and vectors (SVG) alongside raster images.

ChatGPT Images 2.6 at a Glance

Spec	ChatGPT Images 2.6
Developer	OpenAI
Type	Proprietary image generation and editing model
Release date	November 4, 2026
API models	GPT-Image-2.6 Flare, GPT-Image-2.6 Sunburst
Input modality	Text and images (including sketches and markup layers)
Output formats	Raster images, animated GIF, SVG
Speed	50% faster than prior version (OpenAI); Flare as little as 6 seconds at low quality
Quality settings	low, medium, high, xhigh, max (per third-party docs reviews)
Complexity setting	low, medium, high; default high (per third-party docs reviews)
Context window	Not yet confirmed
API pricing	$5 / 1M text input, $8 / 1M image input, $30 / 1M image output tokens (reported)
Batch API discount	Not yet listed for 2.6
Availability	ChatGPT and Codex, all tiers, web/desktop/mobile; OpenAI API
License	Not yet confirmed; no open weights announced
Independent benchmarks	Not yet published

How ChatGPT Images 2.6 Works / What Makes It Different

Five features define this release. Each is named consistently below because these are the terms readers will see across documentation and coverage.

GIF Mode. You can start from a prompt, an uploaded photo or a rough sketch and ask for a short loop. The OpenAI Developer Community announcement confirms you can adjust the speed and edit individual frames. OpenAI's own example prompts request a sprite sheet of exactly 6 frames in a 3-column by 2-row grid, which suggests the model plans animation as a frame grid. Single-source community posts report an API output_format=gif parameter, 2 to 64 frames, and 0.5x to 2x speed control; treat those numbers as unconfirmed.

Saved References. ChatGPT can store reference images of yourself, pets and characters for later conversations. A third-party walkthrough describes naming a saved reference (for example, "Mascot") and calling it with @Mascot in a new prompt. This targets the hardest problem in production image work: keeping the same face or product across separate briefs.

Markup Editing. You draw on an image to show exactly where a change should go. In the API, this reportedly arrives as a transparent PNG layer on the first input image, so you can circle a vase and recolor it while leaving the shelf untouched. Our earlier page on GPT Image 2.6 covers the overlay format in more depth.

SVG Export. The generation endpoint can return a vector file, suited to icons, logos and simple illustrations. Vectors scale from an app icon to a billboard without blurring, and individual paths stay editable in design software.

Complexity Control. In the API, quality governs rendering detail while complexity sets how much the model plans before it renders. Low complexity fits a quick backdrop swap; high fits a poster with many objects and placement rules.

On speed, OpenAI positions Flare for fast drafts and improved layout quality on posters and product imagery. Sunburst trades latency for detail and "even greater image consistency across edits" at medium or high quality.

What You Can Do With ChatGPT Images 2.6

Early use cases cluster around reusable visual identity and motion:

• Brand mascots and campaign loops. OpenAI cites animating a mascot or turning a product shot into an eye-catching loop.
• Reactions, memes and stickers. One Japanese user reported turning a single still image into a character-emotions GIF and flagged sticker mass production as the obvious application.
• Animated storyboards. OpenAI suggests testing character movement before committing to video production.
• Image-first frontend design. One developer describes a Codex workflow where Images 2.6 designs the UI as an image first, and the app is then built using that image as the reference, avoiding the generic dashboard look.
• Vector assets for web and print. SVG export removes the redraw step that used to sit between an AI concept and a production logo or icon.
• Product imagery with stable details. Background swaps that keep a bottle's label, cap and camera angle are the target workload for Sunburst.

A YouTube creator's hands-on video tests GIFs and sprites against Google's Nano Banana. Early user sentiment is mixed: some report fewer noise artifacts, while others say 2.6 looks worse than 2.5 on their own prompts.

How ChatGPT Images 2.6 Compares

The most useful comparison at launch is against its direct predecessor. Rates are identical, but token consumption is not. The figures below come from a third-party analysis of OpenAI's image generation calculator for a 1024x1024 output, excluding input costs.

Metric (1024x1024)	GPT Image 2.5	ChatGPT Images 2.6
Image output rate	$30 / 1M tokens	$30 / 1M tokens
Low quality tokens / cost	196 / ~$0.0059	823 / ~$0.0247
Medium quality tokens / cost	439 / ~$0.0132	1,449 / ~$0.0435
High quality tokens / cost	1,756 / ~$0.0527	1,756 / ~$0.0527
Max quality cost	~$0.2107	~$0.2107
Animated GIF output	No	Yes
SVG export	No	Yes
Batch discount (50%)	Listed	Not yet listed

At high quality and above, ChatGPT Images 2.6 costs the same per image as GPT Image 2.5 (about $0.053 at 1024x1024), while low-quality drafts cost roughly 4.2 times more. For background on the predecessor, see our first look at GPT Image 2.5.

No controlled comparison against Google's Nano Banana 2.1 or ByteDance's Seedream 5 Pro has been published. OpenAI reportedly says Sunburst ranked first on Design Arena, but no Elo score accompanied the claim.

Availability: How to Access ChatGPT Images 2.6

ChatGPT and Codex. Images 2.6 is available globally on all tiers, from Free to Pro, on web, desktop and mobile. The official ChatGPT account confirmed the rollout alongside a sample GIF.

GIF creation is reportedly a gradual rollout. At least one user reported that a GIF request returned a static image instead, so the feature may not appear on every account yet.

Source: @ItsmeAjayKV

OpenAI API. Developers call gpt-image-2.6-flare or gpt-image-2.6-sunburst through the Image API, with rates listed on OpenAI's API pricing page. Teams with cost-sensitive drafting pipelines can keep low-quality drafts on GPT Image 2.5 and reserve 2.6 for high-quality finals, where the two cost the same.

OpenAI also posted a 35-second launch video.

What We Don't Know Yet

• API timing per tier. OpenAI's launch post on X said the two models were "coming" to the API, while the blog says they are offered now. One developer asked when Sunburst reaches all API tiers.
• Independent speed tests. The 50% speedup and six-second Flare figure are vendor claims; no controlled latency measurement exists yet.
• Consistency under stress. Community members proposed tests for product fidelity and looping characters, but no results have been posted.
• GIF limits. Frame counts, maximum duration and resolution for GIF output lack official confirmation.
• ChatGPT plan quotas. Per-tier generation limits are not detailed.
• Batch pricing. The 50% Batch discount does not yet list 2.6 models.

Until independent benchmarks land, the safest reading of ChatGPT Images 2.6 is a real feature expansion with vendor-only performance numbers.

Frequently Asked Questions

What is ChatGPT Images 2.6?

ChatGPT Images 2.6 is OpenAI's image generation and editing update, released on November 4, 2026, that adds animated GIFs, saved reference images, markup-guided edits and SVG export. OpenAI says it generates images 50% faster than before. Developers reach it through two API models, GPT-Image-2.6 Flare and GPT-Image-2.6 Sunburst.

Is ChatGPT Images 2.6 free?

Yes, it is available on all ChatGPT and Codex tiers on web, desktop and mobile, including Free. Per-plan usage limits have not been detailed. API use is billed separately per token.

How much does ChatGPT Images 2.6 cost in the API?

Reported rates are $5 per 1M text input tokens, $8 per 1M image input tokens and $30 per 1M image output tokens, matching GPT Image 2.5. Because 2.6 uses about 4.2 times more tokens at low quality, a low 1024x1024 image costs about $0.025 versus $0.006. At high quality, both cost about $0.053.

What is the difference between GPT-Image-2.6 Flare and Sunburst?

Flare is the speed model and can return an image in as little as six seconds at low quality. Sunburst is the quality model, built for higher detail and stronger consistency across edits at medium or high quality. Both reportedly share the same per-token price.

Can ChatGPT Images 2.6 make GIFs?

Yes, it can turn a prompt or uploaded picture into an animated GIF. In ChatGPT you can adjust speed and edit individual frames. GIF creation is reportedly rolling out gradually, so some accounts may not see it yet.

ChatGPT Images 2.6 vs GPT Image 2.5: what changed?

Version 2.6 adds GIF output, SVG export, saved references, API markup editing and a complexity setting, and OpenAI says it is 50% faster. Per-token rates are unchanged. Token consumption rises at low and medium quality, which raises per-image cost at those settings.

Is ChatGPT Images 2.6 open source?

No, based on available information. OpenAI offers it only through ChatGPT, Codex and its paid API. No weights or self-hosting license have been announced.

What to Watch Next

Three signals will reshape this page. First, independent leaderboard scores for Flare and Sunburst, which would test OpenAI's Design Arena claim. Second, Batch API listing for the 2.6 models, which would cut high-volume costs by 50%. Third, controlled consistency tests on saved references and looping GIFs; for the competitive picture, our Nano Banana 2.1 vs GPT Image 2.5 comparison shows where the baseline sits today.

Try related models on Kie.ai: GPT Image 2.5, Nano Banana 2.1 and Seedream 5.0 Pro.