Skip to main content

China's DeepSeek releases long-awaited new AI model

Chinese startup DeepSeek released a new artificial intelligence model with “drastically reduced” costs on Friday, more than a year after it stunned the world with a low-cost reasoning model that matched the capabilities of US rivals.

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.

Hangzhou-based DeepSeek burst onto the scene in January last year with a generative AI chatbot, powered by its R1 reasoning model, that upended assumptions of US dominance in the strategic sector.

The new version, DeepSeek-V4, “features an ultra-long context of one million words”, the company said in a statement on social media platform WeChat, hailing it as “world-leading… with drastically reduced compute (and) memory costs” in a separate announcement on X.

The model’s context length, which determines how much input a model is able to absorb to help it complete tasks, “(achieves) leadership in both domestic and open-source fields across agent capabilities, world knowledge, and reasoning performance”, the WeChat statement said.

A “preview version” of the open source model is now available, the company said.

Experts say V4’s release marks an “inflection point” in terms of hardware 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,” Zhang Yi, the founder of tech research firm iiMedia, told AFP.

“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 said.

The new V4 is released as two versions, DeepSeek-V4-Pro and DeepSeek-V4-Flash, with the latter being “a more efficient and economical choice” because it has smaller parameters.

‘Sputnik moment’

V4-Pro has 1.6 trillion parameters while the V4-Flash has 284 billion parameters, which refine models’ decision-making ability.

The model has also been “optimised” for popular AI Agent products such as Claude Code, OpenClaw, OpenCode, and CodeBuddy, the DeepSeek statement said.

“In 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’s) Gemini-Pro-3.1,” the statement added.

Last year’s so-called “DeepSeek shock” sparked a sell-off of AI-related shares and a reckoning on business strategy in what was also described as a “Sputnik moment” for the industry.

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.

However, its sudden popularity raised questions over data privacy and censorship, with the chatbot often refusing to answer questions on sensitive topics such as the 1989 Tiananmen crackdown.

At home, DeepSeek’s AI tools have been widely adopted by Chinese municipalities and healthcare institutions as well as the financial sector and other businesses.

This has been partly driven by DeepSeek’s decision to make its systems open source, with their inner workings public — in contrast to the proprietary models sold by OpenAI and other Western rivals.

But the White House has accused Chinese firms of vying to “steal” American technology, 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,” Trump’s science and technology chief advisor Michael Kratsios said in a post on X.

Distillation is a common practice within AI development, often used by companies to create cheaper, smaller versions of their own models.

DeepSeek’s Friday announcement also came as Meta said it planned to cut a tenth of its staff as it looks for productivity gains from the rest of the workforce while investing heavily in artificial intelligence. Reports said Microsoft was also looking to trim its ranks.



from Dawn - Home https://ift.tt/agcL8up

Comments

Popular posts from this blog

AI deepfakes: The latest weapon in Pakistan’s disinfo wars

I have spent the past few weeks staring at what seem to be digital ghosts; they look human, speak the way we do, and cause anger and despair. But they are not real. They are machines wearing human skin. Something is shifting in Pakistan’s social media landscape. It is not merely a rumour. It is not the old cycle of misinformation. It feels darker. Artificial intelligence is reshaping reality frame by frame until the internet resembles a hall of mirrors. For me, it began on November 8, when an account called ‘PakVocals’ posted a video on X that claimed to show journalist Benazir Shah dancing in a nightclub. The caption was cruel. It tried to mock her by using derogatory comments against her professional credibility. At the time of writing this piece, that video had garnered more than half a million views. It did exactly what it was designed to do: turn a journalist into a target. Mission accomplished! To most viewers, the clip probably passed as real, but for me, something felt off....

Ailing Pope Francis to embark on Asia trip, his longest ever, in September

Pope Francis will travel to Indonesia, Papua New Guinea, Timor-Leste and Singapore from September 2-13, the Vatican said on Friday, announcing his first overseas trip of the year and the longest of his 11-year papacy. The Asia trip has been on the papal agenda for some time, but there had been doubts on whether the 87-year-old pontiff would embark on it given his increasing frailty, with a record of skipping engagements due to health problems. His last international journey was a two-day stay in Marseille, France in September. In November, he pulled out of a trip to the COP28 climate conference in Dubai because of a lung inflammation . Francis is now scheduled to be in Jakarta between Sept 3-6, Port Moresby and Vanimo between Sept 6-9, Dili September. 9-11 and Singapore Sept 11-13, his spokesman said in a statement. Vietnam, which had been suggested by the pope and Vatican officials as a possible further destination during the nearly two-week long Asia trip, was not mentioned. In ...

India raises retail fuel prices for first time since Mideast war began

India’s state-run fuel retailers have raised petrol and diesel prices for the first time in four years by 3 Indian rupees ($0.03) per litre, or more than three per cent, according to dealers, to recoup some of the losses incurred due to higher global crude oil prices. India — the world’s third-biggest oil importer and consumer — is one of the last major economies to raise retail fuel prices following the disruption to shipping through the Strait of Hormuz by the war started by US-Israeli attacks on Iran. State-run Indian Oil Corp, Hindustan Petroleum Corp and Bharat Petroleum Corp, which together control more than 90pc of India’s 103,000 fuel stations, tend to set diesel and petrol prices in tandem. A BPCL spokesperson confirmed the price increase at its retail outlets. Indian Oil and HPCL did not immediately respond to a request for comment. Diesel in Delhi will cost 90.67 Indian rupees a litre and petrol 97.77 Indian rupees, reflecting increases of 3.4pc and 3.2pc, respectively, ...