Latest news with #R1-0528
Yahoo
a day ago
- Business
- Yahoo
DeepSeek's R1 Upgrade Nears Top-Tier LLMs
DeepSeek today rolled out DeepSeek-R1-0528, an upgraded version of its R1 large language model that it says now rivals OpenAI's O3 and Google's (NASDAQ:GOOG) Gemini 2.5 Pro. The China-based AI firm credited enhanced post-training algorithmic optimizations and a beefed-up compute pipeline for boosting reasoning accuracy from 70% to 87.5% on complex logic tasks, while cutting hallucination rates and improving vibe coding performance. DeepSeek highlighted benchmark wins in mathematics, programming and general inference, positioning R1-0528 as a peer to leading Western models. This release follows DeepSeek's recent open-source launch of Prover-V2, a specialist reasoning engine, and comes amid a flurry of Chinese AI advancementsAlibaba's (NYSE:BABA) Qwen 3 and Baidu's (NASDAQ:BIDU) Ernie 4.5/X1, both touting hybrid reasoning firepower. DeepSeek argues that its combination of open-development ethos and performance parity gives it a unique edge in global AI research. Investors and partners should care because DeepSeek-R1-0528's near-par with top-tier LLMs could accelerate enterprise deployments in Asia and beyond, drive cloud-compute demand, and intensify competition in the rapidly evolving AI landscape. As Western and Chinese models vie for supremacy, benchmarks like these will shape strategic bets on talent, infrastructure and cross-border AI collaborations. With R1-0528 available now on Hugging Face, markets will watch for adoption by startups and research labs, potential licensing deals, and further advances in DeepSeek's open-source roadmap. This article first appeared on GuruFocus.
Shafaq News
a day ago
- Business
- Shafaq News
DeepSeek unveils upgraded giant-challenging R1 model
Shafaq News/ Chinese AI firm DeepSeek has released an upgraded version of its flagship R1 reasoning model, intensifying competition with US leaders OpenAI and Google. The updated model, R1-0528, significantly enhances performance in complex inference tasks, narrowing the gap with OpenAI's o3 series and Google's Gemini 2.5 Pro, according to a post on the developer platform Hugging Face. While described as a 'minor' version upgrade, R1-0528 introduces substantial improvements in mathematical reasoning, programming, and logical deduction. DeepSeek also reported a 50% reduction in hallucinations—AI-generated false or misleading output—in tasks such as rewriting and summarization. In a WeChat post, the Hangzhou-based firm said the model now excels at generating front-end code, roleplaying, and producing creative writing including essays and novels. 'The model has demonstrated outstanding performance across various benchmark evaluations.' Originally launched in January, R1 quickly went viral, challenging assumptions that advanced AI development requires vast computing infrastructure. Its success triggered responses from Chinese tech giants such as Alibaba and Tencent, both of which released competing models claiming superior performance. DeepSeek also disclosed that it applied a distillation technique—transferring the reasoning methodology from R1-0528—to enhance Alibaba's Qwen 3 8B Base model, boosting its performance by more than 10%. 'We believe the chain-of-thought from DeepSeek-R1-0528 will hold significant importance for both academic research and industrial development focused on small-scale models,' the company added.
TechCrunch
2 days ago
- Business
- TechCrunch
DeepSeek's updated R1 AI model is more censored, test finds
Chinese AI startup DeepSeek's newest AI model, an updated version of the company's R1 reasoning model, achieves impressive scores on benchmarks for coding, math, and general knowledge, nearly surpassing OpenAI's flagship o3. But the upgraded R1, also known as 'R1-0528,' might also be less willing to answer contentious questions, in particular questions about topics the Chinese government considers to be controversial. That's according to testing conducted by the pseudonymous developer behind SpeechMap, a platform to compare how different models treat sensitive and controversial subjects. The developer, who goes by the username 'xlr8harder' on X, claims that R1-0528 is 'substantially' less permissive of contentious free speech topics than previous DeepSeek releases and is 'the most censored DeepSeek model yet for criticism of the Chinese government.' Though apparently this mention of Xianjiang does not indicate that the model is uncensored regarding criticism of China. Indeed, using my old China criticism question set we see the model is also the most censored Deepseek model yet for criticism of the Chinese government. — xlr8harder (@xlr8harder) May 29, 2025 As Wired explained in a piece from January, models in China are required to follow stringent information controls. A 2023 law forbids models from generating content that 'damages the unity of the country and social harmony,' which could be construed as content that counters the government's historical and political narratives. To comply, Chinese startups often censor their models by either using prompt-level filters or fine-tuning them. One study found that DeepSeek's original R1 refuses to answer 85% of questions about subjects deemed by the Chinese government to be politically controversial. According to xlr8harder, R1-0528 censors answers to questions about topics like the internment camps in China's Xinjiang region, where more than a million Uyghur Muslims have been arbitrarily detained. While it sometimes criticizes aspects of Chinese government policy — in xlr8harder's testing, it offered the Xinjiang camps as an example of human rights abuses — the model often gives the Chinese government's official stance when asked questions directly. TechCrunch observed this in our brief testing, as well. DeepSeek's updated R1's answer when asked whether Chinese leader Xi Jinping should be removed. Image Credits:DeepSeek China's openly available AI models, including video-generating models such as Magi-1 and Kling, have attracted criticism in the past for censoring topics sensitive to the Chinese government, such as the Tiananmen Square massacre. In December, Clément Delangue, the CEO of AI dev platform Hugging Face, warned about the unintended consequences of Western companies building on top of well-performing, openly licensed Chinese AI.
Egypt Independent
2 days ago
- Business
- Egypt Independent
China's DeepSeek releases update to AI model that sent US shares tumbling earlier this year
Shanghai Reuters — Chinese artificial intelligence startup DeepSeek released an update to its R1 reasoning model in the early hours of Thursday, stepping up competition with US rivals such as OpenAI. DeepSeek launched R1-0528 on developer platform Hugging Face, but has yet to make an official public announcement. It did not publish a description of the model or comparisons. But the LiveCodeBench leaderboard, a benchmark developed by researchers from UC Berkeley, MIT, and Cornell, ranked DeepSeek's updated R1 reasoning model just slightly behind OpenAI's o4 mini and o3 reasoning models on code generation and ahead of xAI's Grok 3 mini and Alibaba's Qwen 3. Bloomberg earlier reported the update on Wedneday. It said that a DeepSeek representative had told a WeChat group that it had completed what it described as a 'minor trial upgrade' and that users could start testing it. DeepSeek earlier this year upended beliefs that US export controls were holding back China's AI advancements after the startup released AI models that were on a par with or better than industry-leading models in the United States at a fraction of the cost. The launch of R1 in January sent tech shares outside China plummeting in January and challenged the view that scaling AI requires vast computing power and investment. Since R1's release, Chinese tech giants like Alibaba and Tencent have released models claiming to surpass DeepSeek's. Google's Gemini has introduced discounted tiers of access while OpenAI cut prices and released an o3 mini model that relies on less computing power. The company is still widely expected to release R2, a successor to R1. Reuters reported in March, citing sources, that R2's release was initially planned for May. DeepSeek also released an upgrade to its V3 large language model in March.
TECHx
2 days ago
- Business
- TECHx
DeepSeek Releases Update to R1 Model, Nears OpenAI
Home » Latest news » DeepSeek Releases Update to R1 Model, Nears OpenAI Chinese artificial intelligence startup DeepSeek has released an update to its R1 reasoning model, increasing pressure on U.S. rivals like OpenAI. The update, named R1-0528, was launched on developer platform Hugging Face in the early hours of Thursday. However, the company has not made an official public announcement or shared a detailed model description. Despite the quiet release, early results suggest strong performance. The LiveCodeBench leaderboard, a benchmark created by researchers at UC Berkeley, MIT, and Cornell, ranked the updated R1 model just behind OpenAI's o4 mini and o3 in code generation. It also ranked ahead of xAI's Grok 3 mini and Alibaba's Qwen 3. • DeepSeek's R1-0528 is now publicly available on Hugging Face • Ranked closely behind top models from OpenAI on key benchmarks Bloomberg first reported the update on Wednesday. According to the report, a DeepSeek representative revealed in a WeChat group that the company had completed a 'minor trial upgrade' and invited users to begin testing. The move comes as global AI competition heats up. Google's Gemini has introduced discounted access tiers, while OpenAI recently released the o3 Mini, a lightweight model with reduced computing demands. Meanwhile, DeepSeek is widely expected to launch a next-generation model. In March, Reuters reported that R2, a successor to the R1 line, was initially planned for release in May. In the same month, DeepSeek also upgraded its V3 large language model, highlighting its ongoing efforts to keep pace with industry leaders. Though it remains relatively quiet about its releases, DeepSeek's performance on recognized benchmarks signals growing competitiveness in the AI space.
