Kandinsky is the Russian text-to-image model from Sber’s FusionBrain, free to use and strong at understanding Russian prompts.
Kandinsky generates and edits images from text descriptions in Russian and English, and FusionBrain’s platform makes it easy to try, build with and deploy. The model family has grown through generations with improvements in quality, style control and image editing, and it powers creative tools across Sber’s ecosystem. For Russian-speaking creators, Kandinsky offers accessible, free generative imaging with prompt understanding that global tools often lack in Russian.
Kandinsky is free to use in FusionBrain’s web interface with generous limits; API usage for business is priced through Sber’s platform.
Russian-speaking creators, marketers and developers who want free, high-quality image generation with native prompt support.
Kandinsky serves Russian-speaking creators and product teams needing generative imagery.
Open FusionBrain in the browser and generate immediately; developers pick up the API for product integration.
Quality image generation at zero cost for individuals, with Russian prompt support that global tools often lack.
Individuals generate free in FusionBrain immediately; developers integrate through the API, and businesses discuss volume terms with Sber’s platform team.
Free tiers carry daily limits and generation quality trails the top closed models on intricate prompts.
Russian-speaking creators enjoy free generation that understands their prompts without translation.
Individual creators and product teams serving Russian-language audiences.
Free, capable and Russian-native: Kandinsky is the default image generator for the Russian-language internet. Creators lose nothing by trying it first, and developers get a practical API for product features.
Free Russian-native image generation makes Kandinsky the default creative tool for its market. Try it before paying for anything else.
Yes — web generation is free; business API usage is metered through Sber’s developer platform.
Yes — generation and editing capabilities are exposed through FusionBrain’s API for developers.
Against global image models, Kandinsky’s Russian-language prompts, free access and domestic hosting set it apart locally.