DeepSeek closes gap with ChatGPT, Gemini in AI race with R1 upgrade: Cuts hallucinations and improves reasoning
Chinese artificial intelligence firm DeepSeek has introduced an upgraded version of its R1 model, claiming significant improvements in mathematical reasoning, programming, and logical capabilities, while reducing hallucination rates, a common flaw in large language models.
The company revealed details of the new model, DeepSeek-R1-0528, on Thursday via the AI platform Hugging Face, stating that its performance now closely rivals top-tier models such as OpenAI's o3 and Google's Gemini 2.5 Pro. The announcement comes as part of China's ongoing efforts to position itself as a serious contender in the global AI race.
'In terms of comprehensive performance, the updated model is approaching the level of the industry's leading systems,' the startup noted in its Hugging Face release. The upgrade builds upon the success of the original R1, launched in January, which surprised the global AI community by delivering competitive results despite being developed with significantly fewer resources than its Western counterparts.
DeepSeek's rise has captured the imagination of both China's tech sector and political leadership. Following the debut of R1, company founder Liang Wenfeng emerged as a prominent figure within the domestic innovation landscape. His influence was underscored in February when he was invited to a prestigious economic forum hosted by President Xi Jinping. Seated alongside industry titans such as Alibaba's Jack Ma and Tencent's Pony Ma, Liang's presence symbolised the growing confidence in China's home-grown AI talent.
The original R1 model sparked concerns in the West over the pace of China's AI advancements, prompting renewed urgency among major players. Its release also coincided with a brief dip in the share price of Nvidia Corp., the American firm whose chips power much of today's AI infrastructure. However, Nvidia has since recovered, supported by strong data centre investment and optimistic earnings projections.
Based in Hangzhou, DeepSeek is among a new wave of Chinese AI companies striving to challenge the dominance of American tech firms. Its emphasis on cost-efficient yet high-performing models has resonated widely, both at home and abroad.
(With inputs from Bloomberg)
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
New Indian Express
41 minutes ago
- New Indian Express
'Trump said surrender, Modi said yes sir': Rahul targets PM over handling of Indo-Pak standoff
Congress leader Rahul Gandhi on Tuesday accused Prime Minister Narendra Modi of having "surrendered" after a call with US President Donald Trump, amid Washington's claim of brokering a truce between India and Pakistan following Operation the Congress Workers' Convention in Bhopal, the Leader of Opposition in the Lok Sabha said, "A call came from Trump and Narendra ji immediately surrendered; history is witness to this. This is the character of BJP-RSS; they always bow down."Rahul stated that India broke Pakistan in 1971 despite America's a contrast with past Congress leadership, he invoked former PM Indira Gandhi's actions during the 1971 Bangladesh Liberation War, saying, "There was a time in 1971, when America's Seventh Fleet had come, but Indira Gandhi ji had said, 'I will do what I have to do'.""The Congress Party does not surrender. Gandhiji, Nehruji, Sardar Patelji – these are not people who surrender, but people who fight against superpowers," he added.ट्रंप का एक फोन आया और नरेंद्र जी तुरंत surrender हो गए - इतिहास गवाह है, यही BJP-RSS का character है, ये हमेशा झुकते हैं।भारत ने 1971 में अमेरिका की धमकी के बावजूद पाकिस्तान को तोड़ा था। कांग्रेस के बब्बर शेर और शेरनियां Superpowers से लड़ते हैं, कभी झुकते नहीं। Rahul Gandhi (@RahulGandhi) June 3, 2025 Doubling down his attack on the BJP, Rahul said, "I know the BJP-RSS people. If you put even a little pressure and give them just a slight push, they run away out of fear.""Donald Trump called from there and said 'Narender… surrender'. Here Narendra Modi followed Trump's gesture by saying 'Yes Sir'," he Trump has continued to claim that American intervention helped avert a potential nuclear conflict between India and Pakistan, after both countries agreed to a ceasefire following a series of has consistently rejected the claim, stating that neither the US nor its threat of tariffs played any role in the ceasefire talks with Pakistan."From the time Operation Sindoor commenced on May 7 till the time of cessation on May 10, there was conversation between India and the US. But the issue of tariffs never came up in these discussions," MEA spokesperson Randhir Jaiswal said at the weekly press briefing on May Affairs Minister S Jaishankar had also clarified that India and Pakistan were "very, very far away" from a nuclear conflict during their recent an interview to German newspaper FAZ, Jaishankar, in an apparent criticism of the West, said there has been a tendency to link everything in "our part of the world" to a "nuclear problem."The four-day India-Pakistan crisis and why civilian safety can't remain an afterthought
India Today
43 minutes ago
- India Today
Google to launch Pixel 10 series on August 13, Pixel 10 Pro leak offers first look at design
Just days after Google invited fans to its exclusive Pixel Superfans programme event in London, we now have a much clearer picture of when the Pixel 10 series will officially launch. According to multiple reports, including information from trusted leaker MysteryLupin and sources speaking to Android Headlines, Google is reportedly planning to unveil the Pixel 10 lineup on August 13. That's the same date the Pixel 9 series launched last year, and if all goes according to plan, pre-orders will begin on the same day. Shipments are said to begin on August 20, two days earlier than last year's the confirmed event date, we've also got our first look at what seems to be the Pixel 10 Pro. The leaked set of images, reshared by Mystic Leaks, doesn't show a dramatic design overhaul. If these images are the real deal, expect flat sides, a horizontal camera bar at the back, and minimal bezels up front — just like the Pixel 9 Pro. Credit: Mystery Leaks/ Telegram The Pixel 10 series will reportedly include four models this year: the Pixel 10, Pixel 10 Pro, Pixel 10 Pro XL, and the Pixel 10 Pro Fold. One of the biggest expected changes is under the hood — all devices are likely to run on Google's next-gen Tensor G5 chip, said to be manufactured by TSMC. This could bring noticeable improvements in performance and efficiency, especially addressing heating and battery concerns from previous Pixel In terms of software, the new phones are expected to debut with Android 16, packed with AI features across the board. We could see advanced editing tools in Google Photos like Video Generative ML, Speak-to-Tweak voice commands for photo edits, and Sketch-to-Image capabilities that can turn rough drawings into more polished visuals — all powered by on-device also talk of Google finally bringing a triple-camera setup to the base Pixel 10 model — something that's been reserved for the Pro line until now. However, not all changes will be upgrades. Some sensors on the standard model might be similar to what we saw in the Pixel 9a, including a possible downgrade in the ultra-wide camera. Still, Google's image processing software has always played a key role in delivering great photos, so the real-world difference might not be too stark. We'll have to wait and less than two months to go until the rumoured launch, there's still more to learn about the Pixel 10 lineup. Stay tuned to India Today Tech for all the latest on Google's upcoming smartphones.
Time of India
an hour ago
- Time of India
AI 'vibe coding' startups burst onto scene with sky-high valuations
By Anna Tong, Krystal Hu NEW YORK: Two years after the launch of ChatGPT, return on investment in generative AI has been elusive, but one area stands out: software development. So-called code generation or "code-gen" startups are commanding sky-high valuations as corporate boardrooms look to use AI to aid, and sometimes to replace, expensive human software engineers. Cursor , a code generation startup based in San Francisco that can suggest and complete lines of code and write whole sections of code autonomously, raised $900 million at a $10 billion valuation in May from a who's who list of tech investors, including Thrive Capital, Andreessen Horowitz and Accel. Windsurf , a Mountain View-based startup behind the popular AI coding tool Codeium, attracted the attention of ChatGPT maker OpenAI, which is now in talks to acquire the company for $3 billion, sources familiar with the matter told Reuters. Its tool is known for translating plain English commands into code, sometimes called "vibe coding," which allows people with no knowledge of computer languages to write software. OpenAI and Windsurf declined to comment on the acquisition. "AI has automated all the repetitive, tedious work," said Scott Wu, CEO of code gen startup Cognition. "The software engineer's role has already changed dramatically. It's not about memorizing esoteric syntax anymore." Founders of code-gen startups and their investors believe they are in a land grab situation, with a shrinking window to gain a critical mass of users and establish their AI coding tool as the industry standard. But because most are built on AI foundation models developed elsewhere, such as OpenAI, Anthropic, or DeepSeek, their costs per query are also growing, and none are yet profitable. They're also at risk of being disrupted by Google , Microsoft and OpenAI, which all announced new code-gen products in May, and Anthropic is also working on one as well, two sources familiar with the matter told Reuters. The rapid growth of these startups is coming despite competing on big tech's home turf. Microsoft's GitHub Copilot, launched in 2021 and considered code-gen's dominant player, grew to over $500 million in revenue last year, according to a source familiar with the matter. Microsoft declined to comment on GitHub Copilot's revenue. On Microsoft's earnings call in April, the company said the product has over 15 million users. LEARN TO CODE? As AI revolutionizes the industry, many jobs - particularly entry-level coding positions that are more basic and involve repetition - may be eliminated. Signalfire, a VC firm that tracks tech hiring, found that new hires with less than a year of experience fell 24% in 2024, a drop it attributes to tasks once assigned to entry-level software engineers are now being fulfilled in part with AI. Google's CEO also said in April that "well over 30%" of Google's code is now AI-generated, and Amazon CEO Andy Jassy said last year the company had saved "the equivalent of 4,500 developer-years" by using AI. Google and Amazon declined to comment. In May, Microsoft CEO Satya Nadella said at a conference that approximately 20 to 30% of their code is now AI-generated. The same month, the company announced layoffs of 6,000 workers globally, with over 40% of those being software developers in Microsoft's home state, Washington. "We're focused on creating AI that empowers developers to be more productive, creative, and save time," a Microsoft spokesperson said. "This means some roles will change with the revolution of AI, but human intelligence remains at the center of the software development life cycle." MOUNTING LOSSES Some "vibe-coding" platforms already boast substantial annualized revenues. Cursor, with just 60 employees, went from zero to $100 million in recurring revenue by January 2025, less than two years since its launch. Windsurf, founded in 2021, launched its code generation product in November 2024 and is already bringing in $50 million in annualized revenue, according to a source familiar with the company. But both startups operate with negative gross margins, meaning they spend more than they make, according to four investor sources familiar with their operations. "The prices people are paying for coding assistants are going to get more expensive," Quinn Slack, CEO at coding startup Sourcegraph , told Reuters. To make the higher cost an easier pill to swallow for customers, Sourcegraph is now offering a drop-down menu to let users choose which models they want to work with, from open source models such as DeepSeek to the most advanced reasoning models from Anthropic and OpenAI so they can opt for cheaper models for basic questions. Both Cursor and Windsurf are led by recent MIT graduates in their twenties, and exemplify the gold rush era of the AI startup scene. "I haven't seen people working this hard since the first Internet boom," said Martin Casado, a general partner at Andreessen Horowitz, an investor in Anysphere, the company behind Cursor. What's less clear is whether the dozen or so code-gen companies will be able to hang on to their customers as big tech moves in. "In many cases, it's less about who's got the best technology -- it's about who is going to make the best use of that technology, and who's going to be able to sell their products better than others," said Scott Raney, managing director at Redpoint Ventures, whose firm invested in Sourcegraph and Poolside, a software development startup that's building its own AI foundation model. CUSTOM AI MODELS Most of the AI coding startups currently rely on the Claude AI model from Anthropic, which crossed $3 billion in annualized revenue in May in part due to fees paid by code-gen companies. But some startups are attempting to build their own models. In May, Windsurf announced its first in-house AI models that are optimized for software engineering in a bid to control the user experience. Cursor has also hired a team of researchers to pre-train its own large frontier-level models, which could enable the company to not have to pay foundation model companies so much money, according to two sources familiar with the matter. Startups looking to train their own AI coding models face an uphill battle as it could easily cost millions to buy or rent the computing capacity needed to train a large language model. Replit earlier dropped plans to train its own model. Poolside, which has raised more than $600 million to make a coding-specific model, has announced a partnership with Amazon Web Services and is testing with customers, but hasn't made any product generally available yet. Another code gen startup Magic Dev, which raised nearly $500 million since 2023, told investors a frontier-level coding model was coming in summer 2024 but hasn't yet launched a product. Poolside declined to comment. Magic Dev did not respond to a request for comment.
