Newsletter
- ChatGPT is now integrating with the native GIF feature, allowing for the creation of animations through text prompts or photo uploads.
- This feature is driven by a new OpenAI model named Images 2.6, but OpenAI has not officially released this model separately as usual yet.
- This feature is being rolled out gradually, and currently, most users may not be able to use it yet.
The Internet may soon see a continuous stream of AI generating dancing cats GIF, and all of this might be thanks to ChatGPT.
According to The Verge, OpenAI is launching a new feature that allows users to generate GIF animations directly within ChatGPT. It is claimed that this feature is supported by their new image generation model, Images 2.6, however, OpenAI has not officially introduced this model through a regular product announcement to date.
Users can describe an animation in natural language or upload an existing photo, and request that ChatGPT bring it to life. After GIF is generated, users can also adjust the playback speed, view individual frames, edit the animation, and share the final result.
OpenAI demonstrated this capability by using a simple request: "Can you create a gif where a cat dances?" It is reported that the generated interface included control options to adjust the animation speed and to view each frame of the GIF in detail.
We tried to use ChatGPT to generate GIF, but during testing, our account did not yet have this feature available. In our interactions, the model would generate a series of frames, but it would not produce an animation.
It seems that this feature is being rolled out gradually, so users may still need to wait a while before they can get their dancing cats to move.
Images Version 2.6 is also coming soon.
It is reported that this GIF generator is part of the broader Images 2.6 update, which is expected to bring faster image generation speeds. The Verge states that OpenAI is bringing this model to ChatGPT and Codex users, covering web, desktop, and mobile platforms.
Developers will also receive two new API models: GPT-Image-2.6 and Flare, as well as GPT-Image-2.6 and Sunburst. The former is aimed at faster generation, while the latter focuses on higher-quality results. There has not yet been a separate announcement for OpenAI, detailing the capabilities of this new model, and we are also unable to test its performance independently.
This update follows closely on the release of ChatGPT Images 2.5 in September. Images 2.5 brings clearer details, more precise editing capabilities, and a generation speed that is up to 50% faster than its predecessor.
In our previous tests, Images 2.5 was directly compared with Google's Nano Banana 2 across six categories, with each side winning three times. The model of OpenAI was particularly outstanding in terms of illustration quality and improved the issue of excessive sharpening that existed in earlier versions; however, it still made errors when rendering text.
This advancement also occurred in GIF, and it happened several months after OpenAI shut down its independent AI video generation app Sora in March. Compared to the videos generated by Sora, the animations produced by GIF are much simpler, but this new feature allows for some form of dynamic generation capability to be used directly within ChatGPT without the need for a separate app.
OpenAI has not yet provided a specific date for when all ChatGPT accounts will be able to obtain the GIF generation function.
