Sberbank Gigachat Max’s nervous network model is not lower than China Deepseek in Russia tasks. This was an interview with the Vice President of the Sberbank Technological Development block by Vice President of Andrei Belevtsev with the Russian 24 TV channel.
According to him, the bank compared his latest Deepseek work with the Gigachat family. In the answers to many problems that models do not need to reason, Sberber’s neural network has been cope with the banking themes and Russian tasks faster.
“At the same time, the Chinese neural network can show good results in the answers where serious thoughts are necessary. Sberbank is trying to be a leader in the field of artificial intelligence, so we are already developing our AI services. ”
The bank manager believes that artificial intelligence (AI) can be used in almost all fields and sectors of the economy. Most of the time, such vehicles are useful in the software development process.
In savings, it is calculated that AI -based assistants can save 40% of developers’ working time.
Belevtsev also said that deep research technology used in Openai models is an advanced platform to solve complex research problems. According to him, it has the ability to interact with external tools: process the flow of thoughts, open the links, search and interpret the data obtained by interpreting the data. He emphasized that all models of the Gigachat family work with external vehicles due to the function of call functions.
At the same time, the senior executive Sberra said that overrunning the development of AI technologies may create additional difficulties for technological teams. This is due to the innovation of decisions, high resource costs and the rapid development of the industry.
What are you thinking?
Source: Gazeta
Ben Stock is a business analyst and writer for “Social Bites”. He offers insightful articles on the latest business news and developments, providing readers with a comprehensive understanding of the business world.