Meta has released 'Muse Image,' an AI that can create GIF animations, and also announced 'Muse Video,' which can generate videos with sound.

On July 7, 2026, local time, Meta announced its image generation AI ' Muse Image ' and video generation AI ' Muse Video '. Muse Image is already in service, and Muse Video is scheduled to launch soon.
Introducing Muse Image and Muse Video
Muse Image is an AI that can generate images based on text input. Before the final image generation process, it can think and adjust the composition or create materials using tools. For example, it can internally call a QR code creation tool to generate an image that includes a QR code that actually allows access to a target website.

It's also possible to specify which people's faces and other features to include in the image. In the example below, the system creates a customizable image by searching the internet for clothing and bag images based on a photo of a person entered by the user.

Furthermore, you can also create GIF animations. For example, if you input 'Create a 16-frame animation of two characters fighting,' it will calculate the content of each frame and output a 16-frame GIF animation.

Click the image below to see the actual GIF animation.

Muse Image also supports image editing, allowing for highly accurate editing processes such as 'cleaning up old photos,' 'colorizing white photos,' 'changing the weather,' and 'changing the background.'

It also supports image generation using multiple reference images.

In Arena's

Muse Image also ranked second in

Muse Image can be run on the chat service 'Meta AI,' and in the US, it's also available on Instagram Stories. It's also available on WhatsApp in some countries, and will soon be available on Facebook.
Muse Video is an AI that can generate videos with sound. At the time of writing, it is positioned as an early preview version and is expected to be available on Meta AI soon.
Alongside the release of Muse Image, we're sharing an early preview of Muse Video. It offers competitive performance in prompt adherence, visual fidelity, and temporal consistency.
https://t.co/kftRoTgq6W pic.twitter.com/iIFeFGLzoE — AI at Meta (@AIatMeta) July 7, 2026
We're investing in areas with current performance gaps, such as audio-video synchronization and…
Muse Video has already appeared in Arena's ranking of performance for generating videos from text , and it is ranked 3rd, surpassing Alibaba's HappyHorse 1.0 .

Related Posts:
in AI, Posted by log1o_hf






