DeepSeek unveils long-awaited low-cost AI model amid US-China tech tensions
Chinese startup DeepSeek on Friday released a new artificial intelligence model with “drastically reduced” costs, more than a year after it stunned the world with a low-cost reasoning system that matched the capabilities of US rivals. The AI race has intensified rivalry between China and the United States, with the White House on Thursday accusing Chinese entities of a massive effort to steal artificial intelligence technology. Beijing dismissed the allegation as “baseless”. Hangzhou-based DeepSeek burst onto the scene in January last year with a generative AI chatbot powered by its R1 reasoning model, upending assumptions of US dominance in the strategic sector. DeepSeek-V4 “features an ultra-long context”, the company said in a statement on WeChat, hailing it as “world-leading... with drastically reduced compute (and) memory costs” in a separate announcement on X. 🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length. 🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params.… pic.twitter.com/n1AgwMIymu — DeepSeek (@deepseek_ai) April 24, 2026 The model supports a context length of one million “tokens”—small components of text including words or punctuation—putting it on par with Gemini developed by Google. Context length determines how much input a model can absorb to complete tasks. The new system has been released in two versions, DeepSeek-V4-Pro and DeepSeek-V4-Flash, with the latter described as “a more efficient and economical choice” due to its smaller parameters. In terms of “world knowledge”, a benchmark for reasoning, V4-Pro trails only the latest Gemini model, the company said. A “preview version” of the open-source model is now available, though no timeline has been given for a final release. ‘Inflection point’ Experts say V4’s arrival marks an “inflection point” in hardware efficiency and cost. “This addresses the long-standing issues of slower performance and higher costs associated with long context lengths, marking a genuine inflection point for the industry,” said Zhang Yi. “For end users, this will bring widespread, accessible benefits. For instance, if ultra-long context support becomes a standard feature, long-text processing is expected to move beyond high-end research labs and enter mainstream commercial applications,” he added. DeepSeek said V4-Pro has 1.6 trillion parameters, while V4-Flash has 284 billion, enhancing decision-making capabilities. The model has also been “optimised” for AI agent tools such as Claude Code, OpenClaw, OpenCode and CodeBuddy. It can also run on chips manufactured by Huawei, the firm said. Huawei—sanctioned by the US since 2019 over national security concerns—said its Ascend SuperPoD products support the V4 series. DeepSeek’s latest release is a “milestone” for Chinese firms, according to AI industry analyst Max Liu. “It’s a good thing for the entire domestic AI industry. It can provide better models for domestic users and we can now expect a lot more things—more products (and a) more competitive market,” he told AFP. “This is no less shocking than when DeepSeek first came out” if the new model matches leading Western systems, he added. ‘Sputnik moment’ Last year’s so-called “DeepSeek shock” triggered a sell-off in AI-related stocks and a rethink of business strategies, with some describing it as a “Sputnik moment” for the industry. The chatbot performed at a level comparable to ChatGPT and other top US offerings, but required significantly less computing power to develop. However, its rapid rise also raised concerns over data privacy and censorship, with the chatbot often declining to answer sensitive questions such as those related to the 1989 Tiananmen crackdown. DeepSeek’s AI tools have since been widely adopted by Chinese municipalities, healthcare institutions, the financial sector, and other businesses—partly driven by its decision to make systems open source, in contrast to proprietary models from firms like OpenAI. The latest release comes ahead of an expected summit between Donald Trump and Xi Jinping in Beijing next month. “The US has evidence that foreign entities, primarily in China, are running industrial-scale distillation campaigns to steal American AI,” said Michael Kratsios in a post on X. The U.S. has evidence that foreign entities, primarily in China, are running industrial-scale distillation campaigns to steal American AI. We will be taking action to protect American innovation. These foreign entities are using tens of thousands of proxies and jailbreaking… pic.twitter.com/kSp1FReI7J — Director Michael Kratsios (@mkratsios47) April 23, 2026 Distillation is a common AI development technique used to create smaller, more efficient models. China rejected the allegation. “The US claims are entirely baseless,” foreign ministry spokesman Guo Jiakun told a news conference in Beijing. “They are a slanderous smear against the achievements of China’s artificial intelligence industry.”
Chinese startup DeepSeek on Friday released a new artificial intelligence model with “drastically reduced” costs, more than a year after it stunned the world with a low-cost reasoning system that matched the capabilities of US rivals. The AI race has intensified rivalry between China and the United States, with the White House on Thursday accusing Chinese entities of a massive effort to steal artificial intelligence technology. Beijing dismissed the allegation as “baseless”. Hangzhou-based DeepSeek burst onto the scene in January last year with a generative AI chatbot powered by its R1 reasoning model, upending assumptions of US dominance in the strategic sector. DeepSeek-V4 “features an ultra-long context”, the company said in a statement on WeChat, hailing it as “world-leading... with drastically reduced compute (and) memory costs” in a separate announcement on X. 🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length. 🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params.… pic.twitter.com/n1AgwMIymu — DeepSeek (@deepseek_ai) April 24, 2026 The model supports a context length of one million “tokens”—small components of text including words or punctuation—putting it on par with Gemini developed by Google. Context length determines how much input a model can absorb to complete tasks. The new system has been released in two versions, DeepSeek-V4-Pro and DeepSeek-V4-Flash, with the latter described as “a more efficient and economical choice” due to its smaller parameters. In terms of “world knowledge”, a benchmark for reasoning, V4-Pro trails only the latest Gemini model, the company said. A “preview version” of the open-source model is now available, though no timeline has been given for a final release. ‘Inflection point’ Experts say V4’s arrival marks an “inflection point” in hardware efficiency and cost. “This addresses the long-standing issues of slower performance and higher costs associated with long context lengths, marking a genuine inflection point for the industry,” said Zhang Yi. “For end users, this will bring widespread, accessible benefits. For instance, if ultra-long context support becomes a standard feature, long-text processing is expected to move beyond high-end research labs and enter mainstream commercial applications,” he added. DeepSeek said V4-Pro has 1.6 trillion parameters, while V4-Flash has 284 billion, enhancing decision-making capabilities. The model has also been “optimised” for AI agent tools such as Claude Code, OpenClaw, OpenCode and CodeBuddy. It can also run on chips manufactured by Huawei, the firm said. Huawei—sanctioned by the US since 2019 over national security concerns—said its Ascend SuperPoD products support the V4 series. DeepSeek’s latest release is a “milestone” for Chinese firms, according to AI industry analyst Max Liu. “It’s a good thing for the entire domestic AI industry. It can provide better models for domestic users and we can now expect a lot more things—more products (and a) more competitive market,” he told AFP. “This is no less shocking than when DeepSeek first came out” if the new model matches leading Western systems, he added. ‘Sputnik moment’ Last year’s so-called “DeepSeek shock” triggered a sell-off in AI-related stocks and a rethink of business strategies, with some describing it as a “Sputnik moment” for the industry. The chatbot performed at a level comparable to ChatGPT and other top US offerings, but required significantly less computing power to develop. However, its rapid rise also raised concerns over data privacy and censorship, with the chatbot often declining to answer sensitive questions such as those related to the 1989 Tiananmen crackdown. DeepSeek’s AI tools have since been widely adopted by Chinese municipalities, healthcare institutions, the financial sector, and other businesses—partly driven by its decision to make systems open source, in contrast to proprietary models from firms like OpenAI. The latest release comes ahead of an expected summit between Donald Trump and Xi Jinping in Beijing next month. “The US has evidence that foreign entities, primarily in China, are running industrial-scale distillation campaigns to steal American AI,” said Michael Kratsios in a post on X. The U.S. has evidence that foreign entities, primarily in China, are running industrial-scale distillation campaigns to steal American AI. We will be taking action to protect American innovation. These foreign entities are using tens of thousands of proxies and jailbreaking… pic.twitter.com/kSp1FReI7J — Director Michael Kratsios (@mkratsios47) April 23, 2026 Distillation is a common AI development technique used to create smaller, more efficient models. China rejected the allegation. “The US claims are entirely baseless,” foreign ministry spokesman Guo Jiakun told a news conference in Beijing. “They are a slanderous smear against the achievements of China’s artificial intelligence industry.”