Chinese startup DeepSeek released a new artificial intelligence model with \u201cdrastically reduced\u201d costs Friday, more than a year after it stunned the world with a low-cost reasoning model that matched the capabilities of US rivals.<\/p>\n
The AI race has intensified the rivalry between China and the United States, and the White House on Thursday accused Chinese entities of a massive effort to steal artificial intelligence technology.<\/p>\n
<\/p>\n
Hangzhou-based DeepSeek burst onto the scene in January last year with a generative AI chatbot, powered by its R1 reasoning model, that up-ended assumptions of US dominance in the strategic sector.<\/p>\n
The new version, DeepSeek-V4, \u201cfeatures an ultra-long context of one million words\u201d, the company said in a statement on social media platform WeChat, hailing it as \u201cworld-leading \u2026 with drastically reduced compute (and) memory costs\u201d in a separate announcement on X.<\/p>\n
MORE: \u2018Don\u2019t get left behind\u2019: 5 AI skills to learn right now<\/a><\/p>\n The model\u2019s context length, which determines how much input a model is able to absorb to help it complete tasks, \u201c(achieves) leadership in both domestic and open-source fields across agent capabilities, world knowledge, and reasoning performance\u201d, the WeChat statement said.<\/p>\n A \u201cpreview version\u201d of the open source model is now available, the company said.<\/p>\n Experts say V4\u2019s release marks an \u201cinflection point\u201d in terms of hardware and cost.<\/p>\n \u201cThis addresses the longstanding issues of slower performance and higher costs associated with long context lengths, marking a genuine inflection point for the industry,\u201d Zhang Yi, the founder of tech research firm iiMedia, told AFP.<\/p>\n \u201cFor 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,\u201d he said.<\/p>\n The new V4 is released as two versions, DeepSeek-V4-Pro and DeepSeek-V4-Flash, with the latter being \u201ca more efficient and economical choice\u201d because it has smaller parameters.<\/p>\n \u2018Sputnik moment\u2019<\/p>\n V4-Pro has 1.6 trillion parameters while the V4-Flash has 284 billion parameters, which refine models\u2019 decision-making ability.<\/p>\n The model has also been \u201coptimised\u201d for popular AI Agent products such as Claude Code, OpenClaw, OpenCode and CodeBuddy, the DeepSeek statement said.<\/p>\n \u201cIn world knowledge benchmarks, DeepSeek-V4-Pro significantly leads other open-source models and is only slightly outperformed by the top-tier closed-source model, (Google\u2019s) Gemini-Pro-3.1,\u201d the statement added.<\/p>\n Last year\u2019s so-called \u201cDeepSeek shock\u201d sparked a sell-off of AI-related shares and a reckoning on business strategy in what was also described as a \u201cSputnik moment\u201d for the industry.<\/p>\n The chatbot performed at a similar level to ChatGPT and other top American offerings, but the company said it had taken significantly less computing power to develop.<\/p>\n