(March 25): DeepSeek released updates to its V3 model that promise to deliver better programming capabilities, underscoring the Chinese AI start-up’s intent to remain a step ahead of competitors.

The V3-0324 update — posted on Hugging Face this week without a formal announcement — claims to address real-world challenges while setting benchmarks for accuracy and efficiency.

In January, DeepSeek surged past ChatGPT to become the most popular free app on Apple’s US app store. Its achievements, including an R1 model that seemingly performed as well as OpenAI’s best, stunned the industry and sparked a sell-off in US markets. 

The V3 is an older DeepSeek platform. The start-up’s AI services have ignited a debate about whether cutting-edge platforms can be built for far less than the billions that US firms are pouring into data centre construction.

Read also:
Alibaba’s Tsai warns of a ‘bubble’ in AI data centre build-out
DeepSeek narrows AI gap with US to three months, ex-Google China head says

Uploaded by Tham Yek Lee