Chinese AI startup DeepSeek has released an update to its V3 model that improves programming capabilities, demonstrating the company's desire to maintain a lead over its competitors, Bloomberg reports.
Update V3-0324, released this week on Hugging Face without an official announcement, aims to solve practical problems and set new standards for accuracy and performance.
In January, DeepSeek overtook ChatGPT to become the most popular free app on the US App Store. Its R1 model performed on par with OpenAI’s best, surprising the market and sending US stocks plunging.
The V3 model is not DeepSeek's newest development, but the company's success has sparked a discussion about whether it is possible to create powerful AI systems at lower costs than what American technology companies invest in data centers.