New GigaChat Capabilities: Five Minutes of Audio Instead of 20 Pages of Text
Sber’s GigaChat has introduced a new podcast‑generation feature that can turn long, dense documents into a lively, five‑minute audio dialogue, reshaping how people consume information across work, education, and media.
How It Works
Users can upload a question, a document, or a link, and the model extracts key ideas, presenting them as a short audio dialogue. Podcasts can be generated directly via the ‘headphones’ icon in the chat interface or through the dedicated Podcast module.
Users may choose from six voices, configure tone, and set duration up to 10 minutes. Later this year, podcasts can be played on Sber smart speakers or shared to the Zvuk music service.
Reaching a Broader Audience
The feature accelerates information processing: twenty pages of text can become five minutes of conversation.
Professionals can replace long reports with concise audio summaries; students can revise course materials on the go; content creators can generate audio without studio resources. Sber notes that the goal is to make dense information accessible to wider audiences through a universal audio format.
Sber as a Digital Platform
The new function pushes Russia’s IT sector closer to deep integration of generative AI into everyday consumer services.
It expands GigaChat’s potential as a core component of Sber’s ecosystem and supports the country’s broader technological sovereignty in AI. The feature also aligns with modern content‑consumption trends: many people prefer listening over reading. Integrating GigaChat’s podcast tool into educational platforms could help learners convert textbooks into more memorable audio.
Businesses could transform reports into internal audio briefings. Future upgrades may include smart‑speaker integration, multimodal features, and possibly video‑podcasts.
Competition and Russia’s Advantage
Text‑to‑audio conversion is becoming a global trend. One of the latest speech‑synthesis developments — MoonCast (2025) — generates natural‑sounding, podcast‑style voices from text sources using ‘invisible narrators.’
Sber has comparable tools, including SaluteSpeech, which reduces production costs by eliminating the need for human voice actors. The GigaChat podcast launch significantly strengthens Russia’s generative‑AI landscape, particularly in audio.
In coming years, competition will intensify as global players add similar features. The winning platforms will be those that produce natural, emotional, customizable audio indistinguishable from human narration — a standard Sber claims to have reached.
