China's DeepSeek unveils new flagship AI model a year after breakthrough
Sign up now: Get ST's newsletters delivered to your inbox
DeepSeek rolled out the V4 Flash and V4 Pro series, touting top-tier performance in coding benchmarks and big advancements in reasoning and agentic tasks.
PHOTO: REUTERS
Hong Kong - DeepSeek rolled out preview versions of a new flagship artificial intelligence model a year after upending Silicon Valley, calling it the most powerful open source platform in a challenge to rivals from OpenAI to Anthropic.
The Chinese start-up unveiled the V4 Flash and V4 Pro series, touting top-tier performance in coding benchmarks and big advancements in reasoning and agentic tasks. They come with architecture upgrades and optimisation improvements, the start-up said on AI community platform Hugging Face.
DeepSeek singled out a technique it dubbed Hybrid Attention Architecture, which it said improves the ability of an AI platform to remember queries across long conversations. It also pushed the one million-token context window – a leap that allows entire codebases or long documents to be sent as a single prompt.
The V4 arrives more than a year after DeepSeek ignited a trillion-dollar stock market rout with the release of the R1, that rivalled the performance of cutting-edge AI systems from companies like OpenAI but was purportedly built for a fraction of the cost.
Chinese chipmakers rallied on April 24 as investors bet the new model will support demand for local chips.
In a post on WeChat, DeepSeek said the service capacity for the V4 Pro series is extremely limited, due to a computing crunch. The start-up however expects pricing for the model to drop significantly after computing clusters powered by Huawei Technologies’ Ascend 950 chips launch in the second half of 2026.
DeepSeek is currently in talks with Tencent Holdings and Alibaba Group Holding for its first funding round.
The new series is a big step forward in the scale and efficiency that has defined DeepSeek’s rise and put enormous competitive pressure on rivals. In the wake of the R1, tech firms and investors began rethinking the wisdom of pouring billions of dollars into AI development. Those outlays have since sprung back, as American technology giants are projected to invest around US$650 billion (S$830 billion) in 2026 on AI infrastructure and data centres.
DeepSeek touted the new model’s superior performance to the likes of OpenAI’s GPT-5.2 on standard benchmarks, but conceded the V4 trails state-of-the-art models by about three to six months.
Still, DeepSeek emphasised that it is not pushing solely raw capability but also fundamentally lowering costs. The V4 is designed to be deployed on cheaper infrastructure. BLOOMBERG
