Chinese startup DeepSeek, which has managed to attract considerable attention in recent weeks, has announced that it is restricting access to its API service due to a lack of server capacity.
The company reports that it has suspended the ability to replenish API credits for customers to avoid affecting its services. At the same time, credits already purchased will remain with users and will not be canceled.
"The existing recharge amount can continue to be used, please understand!" DeepSeek said.
- Its low prices have caused concern among some competitors, but now DeepSeek says that the discounts for access to its model will end on February 8. The chatbot will then cost 2 yuan per million incoming tokens and 8 yuan per million outgoing tokens.
- And when the inference model goes live, the fee will be 4 yuan per million incoming tokens and 16 yuan per million outgoing tokens.
- DeepSeek unveiled its R-1 AI model in late January, its app became the most popular in the App Store, and the company caused significant fluctuations on Wall Street, leading to losses among tech companies and billionaires.
- However, along with its enormous popularity, the Chinese company's AI is also treated with caution, for the same reasons as the popular Chinese app TikTok, which is "struggling" in the United States.
Recently, Wiz, an American cybersecurity company, revealed leaked correspondence with the DeepSeek AI bot and other confidential data in the public domain. And a number of countries have banned its use in government agencies or in general, fearing a threat to national security. Today, for example, the Dutch government banned civil servants from using a Chinese chatbot because it believes it can be used for espionage. Earlier, Italy blocked access to the chatbot.
