China’s DeepSeek has improved the R1 AI model in Openai competition

DeepSeek, which causes shock waves in the markets this year, left an improved version of the artificial intelligence justification model.

The company did not give an official announcement, but the Improvement of DeepSeek R1 was released on the face of AI model depot hugging.

DeepSEEK has risen to fame this year after free, open source R1 Justification Model Outported Offers From competitors including meta and Openai. The cheapest price and short term of Azerbaijan shocked global markets, the US technology giants have increased over infrastructure, and the AI ​​Stalwart is a billions of dollars worth of large US technological shares such as NVIDIA. These companies have been widely healed since today.

As with the debut of DeepSseek R1, the improved model was also broadcast with small fans. The EU’s step-by-step is a reasonable model that can perform more complex tasks through logical thinking process.

The Improved DeepSeek R1 model is only behind a site on Openai’s O4-Mini and O3-Mini and Livecodebeng, against models against different sizes.

“Deepseek’s latest update, more powerful in math and code, more powerful in highest levels,” Ordinary YAKEFU, EU researcher AI researcher told CNBC.

The improved model “has a great development in the field of reducing hallucinion,” he said. Hallucination belongs to AI, which provides incorrect information.

Deepseleek has become a poster in China’s artificial intelligence, despite the attempts to restrict access to US chips and other technologies. This month, Chinese technology giants Baidu and Tencent revealed how they were Make AI models more efficient Dealing with the export of US semiconductor exports.

Jensen Huang, CEO Nvidiadesigning graphic processing units to train giant AI models, We flushed export departments on Wednesday.

“The United States justifies its policy on the assumption of China’s AI plugs,” Huang said. “This hypothesis was always suspicious and now it’s clearly wrong.”

“The question is whether China is AI,” Huang said. “It’s already.”