The catalog
* FAMILY · sber/kandinsky5

Kandinsky 5 Pro Fast Image to Video

Kandinsky 5.0 Pro is a diffusion-based image-to-video model built for fast generation with smooth motion and visually consistent results across the clip.

1 variantssber provider
* ABOUT KANDINSKY 5

Working with Kandinsky 5

Kandinsky 5.0 Pro is a single diffusion-based image-to-video model from Sber. It prioritises generation speed while keeping motion smooth and the visuals consistent from the first frame to the last.

When to choose Kandinsky 5

Choose Kandinsky 5 for a fast, straightforward animation of a still with no tier decisions. Kling, Veo and Seedance offer more control, audio and longer durations when the shot needs them.

* FREQUENTLY ASKED

About Kandinsky 5

What input does Kandinsky 5 need?
An image. It is image-to-video only, so there is no text-only generation route in this family.
Is it fast?
Speed is its stated design goal, alongside smooth motion and visual consistency across the generated clip.
Does it generate audio?
No. For a clip with sound from the same call, use a family with native audio such as Veo 3 or LTX 2.5.