HONG KONG (AP) \u2014 DeepSeek, the Chinese artificial intelligence startup that shook world markets last year, launched preview versions of its latest major update Friday as the AI rivalry between China and the U.S. heats up.<\/p>\n
DeepSeek\u2019s V4 has been keenly anticipated by users keen to test how it compares to U.S. competitors like OpenAI\u2019s ChatGPT, Anthropic\u2019s Claude and Google\u2019s Gemini. Anthropic and OpenAI have accused DeepSeek of unfairly building its technology off their own.<\/p>\n
Some industry analysts had expected the new model to arrive more than a month earlier at the start of the Lunar New Year.<\/p>\n
DeepSeek says the new V4 open-source models, which include \u201cpro\u201d and \u201cflash\u201d versions, have big improvements in knowledge, reasoning and in their \u201cagentic\u201d capabilities \u2013 the ability to perform complex tasks and workflows autonomously.<\/p>\n
V4 is a successor to V3, an AI model that DeepSeek released in late 2024.<\/p>\n
But it was DeepSeek\u2019s specialized \u201creasoning\u201d AI model, called R1, that took markets by surprise with its release in January 2025. DeepSeek claimed it was more cost-effective than OpenAI\u2019s similar model and it became a symbol of how China was catching up with the U.S. in technological advancements.<\/p>\n
DeepSeek said the \u201cV4 Pro Max\u201d version has \u201csuperior performance\u201d in terms of standard reasoning benchmarks relative to OpenAI\u2019s GPT-5.2 model and Google\u2019s Gemini 3.0-Pro. It falls \u201cmarginally\u201d short of GPT-5.4 and Gemini 3.1-Pro, it said.<\/p>\n
In terms of \u201cagentic\u201d capabilities, the Chinese company said the V4 \u201cpro\u201d version could outperform Claude\u2019s Sonnet 4.5 and approaches the level of Claude’s Opus 4.5 model based on its own evaluation.<\/p>\n
The \u201cflash\u201d version of V4 performs on a par with the \u201cpro\u201d version on simple agent tasks and has reasoning capabilities closely approaching it, DeepSeek said.<\/p>\n
\u201cBased on the benchmark results, it does appear DeepSeek V4 is going to be very competitive against its U.S. rivals,\u201d said Lian Jye Su, chief analyst at the technology research and advisory group Omdia.<\/p>\n
Marina Zhang, an associate professor at the University of Technology Sydney, said DeepSeek’s V4 rollout is as a \u201cpivotal milestone for China\u2019s AI industry\u201d, especially as global competition intensifies in the pursuit of self-reliance in critical technologies.<\/p>\n
DeepSeek offers a free\u2011to\u2011use web and mobile chatbot. Unlike the top models from Anthropic, Google and OpenAI, it describes its technology as \u201copen source\u201d in the way that it enables developers access to modify and build on its core technology.<\/p>\n
Both the V4’s \u201cpro\u201d and \u201cflash\u201d versions have a 1 million token context window, a parameter of how much information an AI model can process and recall, and run on a more efficient basis, the startup said. That is a significant improvement from before, since the V3 supported a 128,000 token context window.<\/p>\n
A report from Microsoft in January showed use of DeepSeek has been gaining ground in many developing nations. <\/p>\n
However, some analysts remain skeptical. Ivan Su, a senior equity analyst at Morningstar, said while V4 is a \u201ccompetent\u201d follow-up, it\u2019s not as big a breakthrough as the rollout of R1. <\/p>\n
\u201cDomestic competition has intensified significantly since R1\u2019s release,\u201d Su said. \u201cAgainst U.S. models, DeepSeek\u2019s own evaluation suggests its capabilities largely match on most fronts, but independent evaluations are needed before final conclusions can be drawn.\u201d<\/p>\n
In February, Anthropic accused DeepSeek and two other China-based AI laboratories of \u201cindustrial-scale campaigns\u201d to \u201cillicitly extract Claude\u2019s capabilities to improve their own models.\u201d It said they did that using a technique called distillation that \u201cinvolves training a less capable model on the outputs of a stronger one.\u201d OpenAI made similar allegations in a letter to U.S. lawmakers.<\/p>\n
This week, Michael Kratsios, chief science and technology adviser to U.S. President Donald Trump, also accused foreign tech companies \u201cprincipally based in China\u201d of distilling leading U.S. AI systems and \u201cexploiting American expertise and innovation.\u201d <\/p>\n
China\u2019s embassy in Washington hit back at the allegations, describing them as \u201cunjustified suppression of Chinese companies by the U.S.\u201d<\/p>\n
___<\/p>\n
O\u2019Brien reported from Providence, Rhode Island.<\/p>\n","protected":false},"excerpt":{"rendered":"HONG KONG (AP) \u2014 DeepSeek, the Chinese artificial intelligence startup that shook world markets last year, launched preview…\n","protected":false},"author":2,"featured_media":15173,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[24,1454,1452,1592,25,1450],"class_list":{"0":"post-15172","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-ai","8":"tag-ai","9":"tag-ap-business","10":"tag-ap-technology","11":"tag-ap-world-news","12":"tag-artificial-intelligence","13":"tag-associated-press"},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/15172","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=15172"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/15172\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/15173"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=15172"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=15172"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=15172"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}