ChatGPT Images 2.6: Create Animated GIFs from Text or Photos

What changed

OpenAI announced that ChatGPT can now create animated GIFs directly within the platform. The feature, which launched on November 4, 2026, is built on the company’s new Images 2.6 model and is available to all ChatGPT and Codex users across web, desktop, and mobile applications.

Users can generate a GIF by entering a text prompt or uploading a photo. Once created, the animation can be customized within ChatGPT. Users can adjust the playback speed and edit individual frames without leaving the interface. OpenAI demonstrated the feature with the prompt “can you make a gif of a cat dancing?” to show how the model maintains character consistency across frames to create a flipbook-style animation.

Alongside GIF support, Images 2.6 also delivers faster overall image generation in ChatGPT. For developers, OpenAI released two new API model variants: GPT-Image-2.6 Flare, designed for efficient generation, and GPT-Image-2.6 Sunburst, built for higher-quality results. Both build on approaches introduced with the earlier Image 2.5 model.

The rollout coincides with continued growth in ChatGPT’s image generation usage. OpenAI reports that users now generate over 4 billion images per week across ChatGPT and the API combined, up from 3 billion per week in September.

Why it matters

For content creators and professionals working with visual assets, this update simplifies workflows. Previously, generating an animated GIF within ChatGPT required exporting static images and using external tools for animation. Now the entire process, from generation to frame-level editing, happens in one interface. This matters for marketing teams producing social media content, designers prototyping animations, and anyone needing quick animated assets without switching between applications.

The feature also demonstrates how image generation tools are becoming capable at tasks traditionally requiring specialized software. The consistency of the model across frames indicates it can maintain visual coherence in ways earlier generations struggled with, addressing a technical challenge that had to be solved for GIF creation to work effectively.

For developers using the API, the availability of two distinct model variants provides flexibility. Teams prioritizing speed can use Flare, while those needing higher fidelity can choose Sunburst. This mirrors a common pattern in AI tooling where vendors offer efficiency and quality as competing priorities.

The growth in image generation volume also warrants attention. An increase of one billion images per week in two months suggests the expanded feature set is attracting usage, though this data does not clarify whether growth came from new users, increased frequency by existing users, or both.

What to test

Before integrating ChatGPT Images 2.6 GIF creation into a professional workflow, consider these practical evaluations:

• Frame consistency across longer sequences. Try prompts requesting complex movements or detailed scenes. Does the model maintain subject identity and visual coherence across 10, 20, or 30 frames? The cat dancing example demonstrates capability, but real-world scenarios vary in complexity.

• Animation smoothness at different speeds. Generate the same GIF and test it at multiple playback speeds. Does it appear natural at slower speeds, or does frame-to-frame variation become jarring? Does faster playback hide or reveal quality issues?

• Editing capabilities and granularity. When editing individual frames, how much control do you actually have? Can you make meaningful edits to one frame without regenerating the entire sequence? Does frame editing work smoothly on mobile, or is it primarily practical on desktop?

• Prompt specificity and consistency. Test whether vague prompts produce usable results or if detailed instructions are necessary. Does the model respond consistently to identical prompts across multiple generations, or does variation make it unreliable for production work?

• File format and export compatibility. Verify what file formats are available for download and whether resulting files work with the platforms where you plan to share them.

• Speed improvements in practice. OpenAI claims faster image generation overall with Images 2.6. Test this against your actual use cases. Does the speed improvement justify switching from an earlier model if you already use one?

Common questions

Can I create GIFs from existing images?

Yes. ChatGPT Images 2.6 allows you to create a GIF from an uploaded photo, not just from text prompts. You can then adjust the animation speed and edit individual frames.

What platforms support GIF creation?

The feature is available across web, desktop, and mobile versions of ChatGPT and Codex, so you can create and edit GIFs on whichever platform fits your workflow.

Are there API options for GIF generation?

OpenAI released two new API models, GPT-Image-2.6 Flare and Sunburst, which support the new capabilities. Flare is optimized for efficient generation while Sunburst prioritizes quality, allowing developers to choose based on their requirements.

The conclusion

ChatGPT Images 2.6 GIF creation represents a clear expansion of the platform’s capabilities. The native ability to generate animated sequences and edit them frame-by-frame removes friction for anyone who previously exported static images to animate them elsewhere. The maintained character consistency across frames solves a real technical problem that affected earlier approaches.

However, this is a new feature, and real-world performance against diverse use cases remains to be determined. The test checklist above reflects the gap between a successful demo and a feature that works reliably at scale. Professionals should try it on their actual prompts and workflows before relying on it for client deliverables or time-sensitive projects.

Pay attention to user feedback on consistency, frame editing usability, and whether the speed improvements deliver across typical usage patterns. Also monitor whether OpenAI releases additional controls or animation options in future updates, which would indicate the company’s commitment to developing this as a production-ready creative tool rather than an experimental feature.

Related reading

• Soniox STT v6: Agent text input and echo filtering
• Perplexity vs ChatGPT Search: Which AI Search Fits You
• The Preference Problem