Recent announcements from Sber show an expanded access plan for GigaChat, the productive artificial intelligence service. The web version is now available to a broad audience who register with Sber ID, enabling seamless text and image processing for a variety of tasks. This move aims to empower individuals and businesses to experiment with and integrate advanced AI into daily operations, education, and creative workflows.
In dialogue with the media, Sber highlighted a key differentiator: the company hosts publicly accessible pre-trained models such as ruGPT, Kandinsky 2.2, and FRED-T5 on its productive AI platform. This strategy is described as a deliberate step to speed up the growth of the open source ecosystem and accelerate AI technology development across Russia. By making these foundational models available, Sber intends to foster collaboration among developers, researchers, startups, and enterprises, while maintaining a focus on practical usability and responsible AI use.
The executive team underscores that wider adoption of GigaChat and its underlying models will drive ongoing evolution. As more users integrate the service into their business processes, new capabilities are unlocked, enabling more ambitious projects and practical applications across industries, from customer support to data analysis and content creation. This collaborative feedback loop is positioned as a catalyst for innovation and broader access to sophisticated AI tools.
Looking ahead, Sber plans to extend GigaChat’s reach through popular messaging platforms. The service is slated to appear in a bot format on Telegram and VK, making interaction more convenient for users who rely on these channels for daily tasks. Developers and organizations will also be able to access the system via a cloud API, enabling scalable integration into custom software, automation pipelines, and enterprise workflows.
GigaChat already integrates with several Sber products and services, with current deployment in devices such as SberBoom and SberBoom Mini smart speakers. In forthcoming updates, the technology will be embedded more extensively across Sber’s smart device lineup, alongside the Salyut virtual assistant, to provide more seamless, hands-free AI capabilities in homes and workplaces.
The GigaChat model was developed by the SberDevices and Sber AI teams, with support from the AIRI Institute of Artificial Intelligence and a group of industry experts who contribute to translating diverse subject matter into the system. Training relies on the Christofari Neo supercomputer, which supports complex computations and large-scale data processing necessary for robust AI performance.
Safety and responsible AI use are central to the service. GigaChat employs a range of safeguards designed to prevent the misuse of generative models and to align with ethical guidelines. The knowledge base provides detailed explanations of capabilities, practical examples, and guidance on how to interact with the system effectively and safely, ensuring users understand both what the model can do and its limitations.
For users already exploring Sber’s ecosystem, GigaChat presents a cohesive AI experience that fits naturally with smart devices, cloud services, and enterprise tools. The ongoing expansion of model access, device integration, and platform reach signals Sber’s commitment to bringing advanced AI into everyday life while supporting innovation, productivity, and responsible deployment across the market.