Chinese startup says DeepSeek-V4-Pro beats all rival open models for maths and coding.

China’s DeepSeek has unveiled the latest versions of its signature artificial intelligence-powered chatbot, a year after its flagship model sent shockwaves through the global tech scene.

The Chinese start-up launched preview versions of DeepSeek-V4-Pro and DeepSeek-V4-Flash on Friday as it touted its ability to go toe-to-toe with US rivals such as OpenAI and Google.

The “pro” version beats all rival open models for maths and coding, and trails only Google’s Gemini 3.1-Pro for world knowledge, DeepSeek said in an announcement on social media.

The “flash” model has similar reasoning abilities to the “pro” version, while offering faster response times and more cost-effective pricing, the Hangzhou-based startup said.

The release comes after DeepSeek-R1 stunned the tech sector upon its launch in January last year with capabilities broadly comparable with those of ChatGPT and Gemini.

Marc Andreessen, a prominent Silicon Valley venture capitalist with close ties to United States President Donald Trump, hailed the model’s release at the time as “AI’s Sputnik moment”.

The performance of the Chinese-developed model attracted particular attention as its developers claimed to have spent less than $6m on computing costs – a fraction of the multibillion-dollar budgets that are usual in Silicon Valley.

DeepSeek’s arrival on the scene also prompted blowback in some countries amid concerns about data protection and Chinese government censorship.

Multiple US states, Australia, Taiwan, South Korea, Denmark and Italy introduced bans or other restrictions on DeepSeek-R1 shortly after its release, citing privacy and national security concerns.

More to follow …