ChatGPT now makes GIFs: Images 2.6 brings motion, two new models and a pricing warning
Direct answer
On November 4, 2026, OpenAI launched Images 2.6 in ChatGPT, Codex and the API. The visible change is that ChatGPT now generates animated GIFs from a prompt or a photo, with playback-speed control and frame-by-frame editing. The API gets two new models, GPT-Image-2.6 Flare (fast) and GPT-Image-2.6 Sunburst (detail and consistency), plus SVG export, markup-guided editing and a complexity setting. OpenAI says generation is 50% faster; per-token rates did not change.
What actually changed
- GIFs in ChatGPT: a prompt or a photo becomes an animated loop. In the API, GIF is one more output format, with 2 to 64 frames per the company's guide. It works for animating a brand mascot or turning a product shot into an eye-catching loop.
- Two models with different jobs: Flare is tuned for latency and fast iteration (about 6 seconds per image at low quality, according to OpenAI) and Sunburst for detail and for keeping a subject steady across a chain of edits.
- More precise editing: markup-guided editing lets you point at the region to change, and the complexity setting (low, medium, high) controls how much detail the model generates.
- Available on every plan: OpenAI says Images 2.6 is in ChatGPT and Codex on all subscription tiers, including Free, on web, desktop and mobile. The rollout was gradual, and some accounts did not have it on day one.
- The per-token price did not change: $5 per million text-input tokens, $8 per million image-input tokens and $30 per million image-output tokens, with lower cached-input rates. The real cost per image depends on the quality chosen and how many output tokens each image consumes.
What it means for a business that creates content
- The GIF is a marketing format, not a toy: a short loop works on social media, WhatsApp and email, where heavy video does not fit. An animated product shot stands out more than a still photo in the same space.
- Speed is a real cost: if generation is faster, the team iterates more and produces more pieces per hour. The 50% figure is an OpenAI claim with no public benchmark, so measure it with your own prompts before promising it.
- Consistency is what matters for a brand: Sunburst aims to keep the same subject identical across several edits. For a company, that matters more than speed: a mascot or a product that changes shape in every image is useless.
- Check the cost per piece, not per token: billing is per token, not per image, so two images of the same size can cost different amounts depending on the quality chosen. Before automating image production, calculate the cost with your real prompts.
Note: this article summarizes OpenAI's November 4, 2026 announcement, «Images, upgraded — introducing ChatGPT Images 2.6», the API documentation and independent analysis from Apidog, OrcaRouter and other outlets. The 50% faster figure is an OpenAI claim with no public benchmark. The GIF rollout was gradual, and some accounts did not receive it on day one.
Want AI that works for your business?
Message us on WhatsApp and let's talk about automation and AI agents for your business.
WhatsApp inicIARelated articles
- How to use AI in your content without losing your voice
- 30 AI terms every business owner should know (glossary)
- 5 AI trends for business in 2026 and 2027
Frequently asked questions
What is Images 2.6?
OpenAI's image-generation update, launched November 4, 2026 in ChatGPT, Codex and the API. It adds animated GIFs, SVG export, markup-guided editing, a complexity setting and two new API models: Flare and Sunburst.
Can ChatGPT make GIFs now?
Yes. With Images 2.6, ChatGPT turns a prompt or a photo into an animated GIF, with playback-speed control and frame-by-frame editing. It is on every plan, including Free, although the rollout was gradual.
What is the difference between Flare and Sunburst?
Flare is tuned for latency and fast iteration; Sunburst for detail and for keeping a subject steady across a chain of edits. Both carry the same per-token rates.
Did the price of generating images change?
Per-token rates did not change: $5 per million text-input tokens, $8 per million image-input tokens and $30 per million image-output tokens. The cost per image depends on the quality chosen and the tokens each generation consumes.
