The open video model just learned to lip-sync. kandinsky 6.0 video, out oct 11, ships a 29 billion…
the open video model just learned to lip-sync. kandinsky 6.0 video, out oct 11, ships a 29 billion parameter pro that makes five-second clips with synchronized 44 khz audio.
audio and picture no longer render separately.
Context
Kandinsky 6.0 Video is a family of diffusion models for synchronized text-to-audio-video generation: Video Lite at 3 billion parameters and Video Pro at 29 billion. Both generate 5-second clips with synchronized 44 kHz audio, including lip-sync, per the Hugging Face model cards and the GitHub repository.
Kandinsky Lab's site describes Lite as built for speed and Pro as the full-scale model with richer motion and camera control.
The 29 billion parameter Pro, five-second clips, 44 kHz synchronized audio and lip-sync all match the model cards. The Oct 11 release date was not seen in the sources read, so unsupported here, not refuted. The pages read carry no date. Whether the weights are open under a permissive license was also not seen in the sources read, so unsupported here, not refuted. 'Audio and picture no longer render separately' is the author's opinion.
Related work
- Hugging Face: Kandinsky-6.0-Lite-5s-Diffusers ↗The 3 billion parameter Lite model.
- Hugging Face: Kandinsky-6.0-Pro-pretrain-5s-Diffusers ↗The pretrain Pro checkpoint.
- Kandinsky Lab video page ↗Lite and Pro descriptions.
Watch next
- Read the Kandinsky 6.0 license text on the model card.
Sources
- Hugging Face: Kandinsky-6.0-Pro-5s-Diffusershuggingface.co
- Kandinsky Lab: Kandinsky 6.0 Videokandinskylab.ai
- GitHub: kandinskylab/kandinsky-6github.com
Provenance
The note above is reproduced unedited from the original post, first published on Threads on 11 October 2026 at 19:10 IST. Sources are the papers and datasets the note draws on.
View the original post ↗Embed this note
More notes
The air is now being asked to keep its own ledger
the air is now being asked to keep its own ledger: ecmwf’s aifs compo becomes the first ai model to forecast atmospheric composition globally every three hours, cleanair simulates 365 days of pm2.5 over china in ten seconds, and a unified framework maps six pollutants at one kilometer across the whole country. the air now files its own composition report.
read the note →The current is now being asked to draw its own map
the current is now being asked to draw its own map: china’s langya 2.0 predicts six ocean phenomena including internal waves and mesoscale eddies, a deep net called wenhai resolves eddies globally with air sea flux formulas built in, and scripps infers surface currents from the way temperature patterns deform in satellite images. the ocean now files its own circulation report.
read the note →The soil is now being asked to report its own carbon
the soil is now being asked to report its own carbon: a nix color sensor paired with generative data augmentation predicts soil organic carbon without a lab, random forest drives 74 percent of soil health mapping studies, and sentinel 2 tracks five year carbon change across france and italy from 922 samples. the dirt now files its own carbon account.
read the note →