Written by Smartech Daily Team
This article has been originally published on Smartech Daily and republished at Dataconomy with permission.
Sber has introduced a major update to its AI assistant GigaChat, which is now powered by the new flagship model GigaChat Ultra. The new version empowers the AI assistant to retain facts about users for more personalized communication and solutions, independently search for information online, and deliver text responses in half the time.
The new model gives opportunities for both end users and developers—enabling developers to build applied AI products and services on GigaChat Ultra. Users can run code in the interface and get accurate answers about their capabilities based on current documentation.
Anton Frolov, senior vice president, head of Generative AI Development, Sberbank:
“We are taking a step from being just an answer-giving tool to becoming a multi-agent AI assistant. But our goal goes beyond that: we are creating a future where traditional mobile apps give way to neural-network-based interfaces. Needed features will appear upon request, making navigation through the digital world seamless. GigaChat Ultra is one of the world’s largest models fully developed and trained in Russia. It remembers your preferences, works faster, understands tasks more deeply, and provides higher-quality recommendations. We are erasing the last boundaries between people and machines.”
Long-term memory
One key innovation lies in long-term memory. Unlike contextual (short-term) memory, which resets after each conversation, GigaChat retains user-specific facts across sessions and uses them in future interactions.
Here’s what exactly GigaChat remembers:
- hobbies, tastes, and interests;
- profession, education, life goals, and habits;
- personal data—only to the extent shared by users themselves;
- information about family members and pets.
The system automatically identifies significant facts without overloading memory with trivialities such as short-term plans or widely known general knowledge. All data is stored in a unified profile synchronized between web versions, mobile applications, and Telegram bot via the Sber ID sign-in. Users have full control over this feature: memory can be enabled or disabled anytime in settings.
Response generation speed doubled
GigaChat generates text responses twice as fast as Sber’s previous flagship model. This means users see replies almost instantly, even for complex queries requiring detailed reasoning.
This increase in speed was achieved thanks to the Mixture of Experts (MoE) architecture. The model acts like a team of specialized experts, each handling specific types of tasks. Only relevant “experts” respond to any given query rather than the entire model working simultaneously.
Real-time conversation mode
GigaChat now autonomously connects to internet searches for real-time updates, eliminating the need for users to manually enable this option. This ensures accurate responses when discussing recent news items, stock quotes, and other dynamically changing data. Search functionality includes a dedicated rephraser—a system that reformulates user queries to enhance relevance and improve final response quality.
Voice mode now supports online search. Conversations are fully interactive: users can interrupt, clarify, or change topics instantly, with no delay in context switching. A complete transcript is saved after each session.
Self-awareness: GigaChat knows everything about itself
A self-awareness mechanism has been implemented in GigaChat, enabling the model to provide correct answers regarding its own characteristics. When responding to these kinds of questions, the model refers to up-to-date documentation describing its current version, supported functionalities, limitations, and behavioral peculiarities. This eliminates typical issues common among language models, such as providing incorrect or outdated information about their abilities—for example, falsely claiming nonexistent features or failing to recognize existing ones.
Code interpreter: GigaChat as an analytical environment
An integrated code interpreter transforms GigaChat into an isolated execution environment for running software code right inside the assistant’s interface. Before introducing this function, the model could merely write code and display it to users; executing and testing results required external tools. Now, GigaChat generates and instantly executes code within a secure sandbox, isolated from the user’s system.
The interpreter supports uploaded files, performs advanced numerical calculations, validates data structures, and creates graphs and charts directly in chats. This makes GigaChat a comprehensive analytical tool suitable for reports, tables, and large datasets.
The training process
The training involved three stages. Initially, the scope of knowledge was expanded by adding academic books, materials related to mathematics and programming, increasing multilingual data volumes—now covering ten languages. In the intermediate stage, specialized skills were enhanced: the code corpus was enlarged, additional data included physics, medicine, finance, records of actual dialogs, and security measures strengthened. Final tuning based on examples (editor texts, dialogs triggering functions, system prompts) ensured stable performance under real-world conditions.
Significant improvements were recorded in open-ended and closed-ended question answering, along with tasks demanding sophisticated logical reasoning. Benchmark tests for Russian-language use demonstrated high levels of grammatical correctness, natural speech flow, readability, and structured responses. Enhancements also extended to practical industry scenarios: the model became more adept at legal, cybersecurity, medical, financial, and trade-related tasks—especially those involving Russian-specific nuances and sectoral terminology. Significant gains in mathematical computation and code generation have widened its applicability in fintech, education, and development.
Flagship model to be released publicly
Sber make the source code and weights of its flagship GigaChat Ultra model freely accessible. According to company experts’ assessments, it already outperforms DeepSeek V3.1, Qwen3-235B and its predecessor GigaChat 2 Max in Russian-language tasks, maths and general reasoning. By releasing the repository, organizations ranging from large banks to small startups will gain the ability to install the neural network within their private environments and adapt it to corporate data, marking a move toward genuine technological sovereignty.
Users can try the updated model free of charge in the web version, Android apps available in RuStore and AppGallery, as well as in the Telegram bot and MAX messenger. To activate voice mode and memory, simply sign in via Sber ID and turn on desired options in profile settings.
