The smartphone screen displays the DeepSeek logo against a backdrop of the Chinese flag.
Sopa Images | Light Rocket | Getty Images
Chinese artificial intelligence startup DeepSeek on Friday released a preview version of its long-awaited V4 large-scale language model, allowing users to test its new features.
The release comes more than a year after the Hangzhou-based company introduced its R1 inference model, which shook up the global technology market with its incredible performance and cost efficiency.
As with previous model releases of DeepSeek, the latest upgrade is open source, allowing developers to download the code, run it locally, and modify it in most cases.
The model is available in both “pro” and “flash” versions depending on size, and DeepSeek claims that V4 achieves strong performance against domestic competitors, especially in agent-based tasks, knowledge processing, and inference.
Neil Shah, vice president of research at Counterpoint Research, told CNBC that “DeepSeek’s V4 preview is serious flex,” offering lower inference costs than previous models.
Inference costs refer to the computational and financial costs of running a trained AI model and producing output.
DeepSeek also said V4 is optimized for use with popular agent tools such as Anthropic’s Claude Code and OpenClaw.
According to Wei Sun, principal AI analyst at Counterpoint, V4’s benchmark profile suggests it can provide “greater agent functionality at a significantly lower cost.”
Will DeepSeek shock the world again?
DeepSeek was founded in 2023 and gained attention in late 2024 with its free and open source V3 model. This model is trained using a less powerful chip and uses OpenAI and google.
A few weeks later, in January 2025, the company released its inference model R1, which achieved similar benchmarks or outperformed many of the world’s leading LLMs.
The R1 model alarmed investors when DeepSeek revealed that it took just two months and less than $6 million to build the model using low-capacity Nvidia chips. This called into question America’s lead in AI and Big Tech’s massive spending on AI infrastructure.
Since then, DeepSeek has released a series of model upgrades, but none have matched the impact of R1.
Ivan Su, senior equity analyst at Morningstar, told CNBC that V4’s debut is unlikely to have as much of an impact on the market as R1 because traders have already priced in the reality that China’s AI will be competitive and cheap.
However, with DeepSeek’s latest positioning, other Chinese open source models are direct competitors, Su said.
“This is a framework that did not exist in R1, and this alone shows how intense domestic competition has become,” he added.
Since the release of R1, DeepSeek has faced increased competition in China’s burgeoning AI space, with players including: alibaba ByteDance will also release new models this year.
Shares of several other Chinese AI companies fell in Hong Kong trading on Friday. Knowledge Atlas Technology, also known as MiniMax and Zhipu, each fell about 8%, while Hangzhou-based developer Manycore Tech fell 9%.
What chip trained V4?
The big question surrounding the release of DeepSeek’s V4 model is which chip was used for its training and support.
Chinese technology giant Huawei on Friday confirmed that its latest AI computing cluster powered by Ascend AI processors can support DeepSeek’s V4 model.
However, it remains unclear how extensively Huawei’s chips were used in training compared to the US AI chip leader’s chips. Nvidia.
The U.S. government’s ever-changing export regulations restrict Chinese developers from directly purchasing Nvidia’s cutting-edge AI chips.
Meanwhile, the Chinese government has stepped up efforts to develop the domestic chip industry, reportedly encouraging Chinese tech companies to adopt domestic alternatives to chipmakers such as Huawei.
Counterpoint’s Wei Sun said the ability of V4 to run natively on local chips could have a huge impact and help the Chinese government further increase its AI sovereignty and further reduce its dependence on Nvidia.
“This will ultimately also accelerate global AI development,” she added.
Shares of the Chinese contract chipmaker rose in Hong Kong after DeepSeek announced its V4 release. SMIC and Huahong Semiconductor They jumped 9% and 15%, respectively.
