Honor is preparing to launch the Honor 600 series next week, and while the hardware specs look promising, the software could be the real star of the show. Building on the foundation of the Honor 400 series, where users have created over 13.4 million seconds of AI video, Honor will introduce an updated version of its AI image-to-video feature with the upcoming devices.
On-device AI for more privacy and speed
Honor’s Image to Video 2.0 uses a unified multimodal video generation model that runs entirely on-device. Unlike cloud-based tools that can be slow or raise privacy concerns, this engine processes everything locally. Additionally, Honor says the workflow is fairly straightforward, making it easier for users to get started.
This feature allows you to provide the AI ​​with one to three reference images and use natural language commands to describe the scene. You can specify specific actions or have the AI ​​fill in the gaps between a start and an end image. The feature will produce a 3 to 8 second clip that is said to appear less like a slideshow and more like a professionally produced video.
How Honor fixes glitches
One of the biggest hurdles for generative videos is visual consistency. When rendering objects from different angles, AI often makes mistakes, resulting in distortion or glitches. Honor claims to have solved this problem with the design of the motif from multiple angles. By understanding the 3D structure of a subject from 2D images, the model remains consistent even with complex camera movements.
For beginners, the Honor 600 series also offers templates for professional filmmaking techniques such as Hitchcock zoom, bullet time and world transformation effects. And to ensure that the tool doesn’t sink deep into the settings, the devices have a dedicated physical AI button.
With the reveal just a week away, Honor is clearly positioning the 600 series as a portable production studio and not just another affordable flagship phone. It remains to be seen how the updated AI Image to Video 2.0 feature performs in the real world.
