BEIJING: The Chinese AI startup DeepSeek has unveiled a significant upgrade to its V3 large language model, ramping up the competition against major U.S. tech companies like OpenAI and Anthropic.
Named DeepSeek-V3-0324, this latest model is accessible via the AI development platform Hugging Face, representing the company’s vigorous effort to carve out its niche in the ever-evolving AI landscape.
This updated model shows remarkable improvements in reasoning and coding skill sets compared to its forerunner. Benchmark tests reveal enhanced performance across various technical metrics available on Hugging Face.
In recent months, DeepSeek has quickly positioned itself as a significant contender in the global AI arena, releasing a succession of models that challenge their Western equivalents while maintaining lower operational costs.
After launching its V3 model in December, the company followed up with its R1 model in January.
This version rephrases the original content while retaining its essential details and context, ensuring it’s unique and suitable for American English.
