- 2Releases tracked
- Nov 3Latest · Kandinsky Avatar 1.0 (S2V 1.0 Lite and Pro)
- 100%Open-weight
- RURussia
Kandinsky Lab is an AI research team of more than 100 researchers, engineers and analysts that develops the open Kandinsky family of generative image and video models. Its lineup covers Kandinsky Image, Kandinsky Video, K-VAE tokenizers for audio, image and video, and Kandinsky WM world models for driving, robotics and physics, with a history dating back to Kandinsky 1.0 in 2021 to 2022. Its tracked release, Kandinsky 6.0 Video (October 2026), open-sourced 3B Lite and 29B Pro diffusion models that generate five-second clips with synchronized audio under the MIT license.
Kandinsky Lab (Sber)
2 releases latest Nov 3, 2026
-
Kandinsky Avatar 1.0 (S2V 1.0 Lite and Pro)
Open speech-to-video avatar models (2B Lite, 21B Pro, plus distilled variants): talking video from one image, a speech track and a prompt. MIT.
-
Kandinsky 6.0 Video (Lite and Pro)
Open-sourced 3B Lite and 29B Pro text/image-to-audio-video diffusion models: 5s clips with synced 44 kHz audio and lip-sync, upscaling to 1080p. MIT.
Kandinsky Lab (Sber) FAQ
- What is Kandinsky Lab (Sber)'s latest AI model?
- Kandinsky Lab (Sber)'s most recent release in the Startrise tracker is Kandinsky Avatar 1.0 (S2V 1.0 Lite and Pro), released Nov 3, 2026.
- How many AI models has Kandinsky Lab (Sber) released?
- Startrise tracks 2 notable Kandinsky Lab (Sber) releases since January 2025, from Kandinsky 6.0 Video (Lite and Pro) (Oct 6, 2026) to Kandinsky Avatar 1.0 (S2V 1.0 Lite and Pro).
- Does Kandinsky Lab (Sber) release open-weight models?
- Yes. 2 of 2 tracked Kandinsky Lab (Sber) releases (100%) ship with downloadable weights; the most recent is Kandinsky Avatar 1.0 (S2V 1.0 Lite and Pro).
- Where is Kandinsky Lab (Sber) based?
- Kandinsky Lab (Sber) is based in Russia.