DeepSeek has recently released an update to its large language model that the company claimed featured “significant improvements” over its predecessor.

The China-based startup appeared to escalate its rivalry with OpenAI and other U.S. artificial intelligence firms, following an earlier release that shook global tech stocks.

On Tuesday, DeepSeek introduced an upgrade to its V3 large language model, DeepSeek-V3-0324, on the AI development platform Hugging Face.

The startup marketed the upgrade as incorporating enhancements in coding and reasoning capabilities compared to its previous V3 model.

On December 26, 2024, DeepSeek released its V3 large language model, which the company claimed was three times speedier than its V2 model.

The model was ranked among the top 10 in Chatbot Arena, a performance platform hosted by the University of California, Berkeley.

In addition to upgrades to front-end web development, Chinese writing proficiency—with new features such as “interactive rewriting”—and Chinese search capabilities—such as “enhanced report analysis,” the most recent update included “significant improvements” across several benchmark tests for the language model, according to DeepSeek.