Until recently, Sber, Yandex and VK actively talked about their achievements in the field of artificial intelligence. Now there are fewer high-profile presentations, and the main work has moved behind closed doors. Russian IT giants are developing their own large language models (LLM), which should become the basis for banks, businesses and government services.
Just a few years ago, the artificial intelligence market resembled a technological race in public, recalls xrust. Companies competed in the number of model parameters, query processing speed, and new neural network functions. Each major development was accompanied by presentations and loud statements.
Today the situation has changed.
Sber, Yandex and VK have become noticeably more careful about their plans. But this silence does not mean slowing down. On the contrary, right now the question is being resolved: who will be able to create a Russian language model capable of working at the level of modern corporate solutions.
From chatbots to full-fledged AI systems
The main goal of companies has changed. A simple assistant who can answer questions or write texts is no longer enough.
Businesses need systems that can analyze documents, help make decisions, work with large amounts of data and perform complex tasks without constant human supervision.
The ability to deploy such models within your own infrastructure is becoming especially important.
For banks, telecom companies and government organizations, this means less dependence on external services, foreign cloud platforms and third-party APIs.
This is why the struggle for your own LLMs is gradually becoming not only a technological, but also a strategic issue.
Sber relies on heavy artillery
Sber remains one of the main participants in the Russian artificial intelligence race.
The company is developing a family of GigaChat models and gradually expanding their capabilities. Unlike the first generations of neural networks, new versions should not just generate responses, but perform more complex tasks — analyze information, work with code and help specialists in various fields.
Sber representatives previously announced the development of models with reasoning capabilities that should bring domestic solutions closer to the modern level of global AI systems.
Source: Reuters reported on Sber's plans to develop large language models with advanced analysis and reasoning capabilities.
Yandex retains its main advantage
Yandex has its own strong foundation — many years of experience in search, machine learning and processing huge volumes of data.
The company is developing the YandexGPT line, integrating artificial intelligence into its services — from search to corporate tools.
The main advantage of Yandex lies not only in the model itself, but also in the accumulated data and experience of working with millions of user requests.
However, this is precisely why the company is forced to act carefully. Any new product is automatically compared with the largest foreign analogues, and errors become noticeable almost instantly.
VK goes through rapid experiments
VK takes a different position.
- Если Вам понравилась статья, рекомендуем почитать
- Hackers infected hundreds of npm packages with ransomware
- WordPress is under attack: critical WP2Shell vulnerability is already being exploited
