{"id":650152,"date":"2025-01-21T00:45:16","date_gmt":"2025-01-20T21:45:16","guid":{"rendered":"https:\/\/en.buradabiliyorum.com\/deepseek-claims-its-reasoning-model-beats-openais-o1-on-certain-benchmarks\/"},"modified":"2025-01-21T00:45:16","modified_gmt":"2025-01-20T21:45:16","slug":"deepseek-claims-its-reasoning-model-beats-openais-o1-on-certain-benchmarks","status":"publish","type":"post","link":"https:\/\/buradabiliyorum.com\/en\/deepseek-claims-its-reasoning-model-beats-openais-o1-on-certain-benchmarks\/","title":{"rendered":"#DeepSeek claims its reasoning model beats OpenAI’s o1 on certain benchmarks"},"content":{"rendered":"
\n

Chinese AI lab DeepSeek has released an open version of DeepSeek-R1, its so-called reasoning model, that it claims performs as well as OpenAI\u2019s o1 on certain AI benchmarks. <\/p>\n

R1 is available from the AI dev platform Hugging Face under an MIT license, meaning it can be used commercially without restrictions. According to DeepSeek, R1 beats o1 on the benchmarks AIME, MATH-500, and SWE-bench Verified. AIME employs other models to evaluate a model\u2019s performance, while MATH-500 is a collection of word problems. SWE-bench Verified, meanwhile, focuses on programming tasks.<\/p>\n

Being a reasoning model, R1 effectively fact-checks itself, which\u00a0helps it to avoid some of the\u00a0pitfalls\u00a0that normally trip<\/a> up models. Reasoning models take a little longer \u2014 usually seconds to minutes longer \u2014 to arrive at solutions compared to a typical nonreasoning model. The upside is that they tend to be more reliable in domains such as physics, science<\/a>, and math.<\/p>\n

R1 contains 671 billion parameters, DeepSeek revealed in a technical report<\/a>. Parameters roughly correspond to a model\u2019s problem-solving skills, and models with more parameters general<\/a>ly perform better than those with fewer parameters. <\/p>\n

671 billion parameters is massive, but DeepSeek also released \u201cdistilled\u201d versions of R1 ranging in size from 1.5 billion parameters to 70 billion parameters. The smallest can run on a laptop. As for the full R1, it requires beefier hardware, but it is<\/em> available through DeepSeek\u2019s API at prices 90%-95% cheaper than OpenAI\u2019s o1. <\/p>\n

There is a downside to R1. Being a Chinese model, it\u2019s subject to\u00a0benchmarking<\/a>\u00a0by China\u2019s internet regulator to ensure that its responses \u201cembody core social<\/a>ist values.\u201d R1 won\u2019t answer questions about Tiananmen Square, for example, or Taiwan\u2019s autonomy.<\/p>\n

\"DeepSeek
R1\u2019s filtering in action. <\/span>Image Credits:<\/strong>DeepSeek<\/span><\/figcaption><\/figure>\n

Many<\/a>\u00a0Chinese AI systems, including other reasoning models,\u00a0decline\u00a0to respond to topics that might raise the ire of regulators in the country, such as speculation about the\u00a0Xi Jinping<\/a>\u00a0regime.<\/p>\n

R1 arrives days after the outgoing Biden administration\u00a0proposed\u00a0harsher<\/a> export rules and restrictions on AI technologies for Chinese ventures. Companies in China were already prevented from buying advanced AI chips, but if the new rules go into effect as written, companies will be faced with stricter caps on both the semiconductor tech and models needed to bootstrap sophisticated AI systems.<\/p>\n

In a policy document last week, OpenAI urged the U.S. government to support the development of U.S. AI, lest Chinese models match or surpass them in capability. In an interview<\/a> with The Information, OpenAI\u2019s VP of policy Chris Lehane singled out High Flyer Capital Management, DeepSeek\u2019s corporate parent, as an organization of particular concern.<\/p>\n

So far, at least three Chinese labs \u2014 DeepSeek, Alibaba, and Kimi<\/a>, which is owned by Chinese unicorn Moonshot AI \u2014 have produced models that they claim rival o1. (Of note, DeepSeek was the first \u2014 it announced a preview of R1 in late November.) In a post<\/a> on X, Dean Ball, an AI researcher at George Mason University, said that the trend suggests Chinese AI labs will continue to be \u201cfast followers.\u201d<\/p>\n

\u201cThe impressive performance of DeepSeek\u2019s distilled models [\u2026] means that very capable reasoners will continue to proliferate widely and be runnable on local hardware,\u201d Ball wrote, \u201cfar from the eyes of any top-down control regime.\u201d<\/p>\n<\/div>\n

If you liked the article, do not forget to share it with your friends. Follow us on\u00a0Google News<\/a><\/span>\u00a0too, click on the star and choose us from your favorites.<\/span><\/strong><\/p><\/blockquote>\n

\n

If you want to read more like this article, you can visit our Technology<\/a><\/span> category.<\/strong><\/p>\n<\/blockquote>\n

Source<\/a><\/span><\/p>\n","protected":false},"excerpt":{"rendered":"

Chinese AI lab DeepSeek has released an open version of DeepSeek-R1, its so-called reasoning model, that it claims performs as well as OpenAI\u2019s o1 on certain AI benchmarks. R1 is available from the AI dev platform Hugging Face under an MIT license, meaning it can be used commercially without restrictions. According to DeepSeek, R1 beats…<\/p>\n","protected":false},"author":1,"featured_media":650153,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"fifu_image_url":"https:\/\/techcrunch.com\/wp-content\/uploads\/2024\/04\/GettyImages-1652364481.jpg?resize=1200,675","fifu_image_alt":"","footnotes":""},"categories":[18],"tags":[77337,153641,153751,153752,153753,147146,153754,141741,153755,153756,151454],"class_list":["post-650152","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-technology","tag-ai","tag-chinese-ai","tag-chinese-labs","tag-deepseek","tag-deepseek-r1","tag-generative-ai","tag-high-flyer","tag-open-ai","tag-r1","tag-reasoning-model","tag-tc"],"_links":{"self":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts\/650152","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/comments?post=650152"}],"version-history":[{"count":0,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts\/650152\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media\/650153"}],"wp:attachment":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media?parent=650152"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/categories?post=650152"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/tags?post=650152"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}