{"id":14997,"date":"2026-02-03T09:42:28","date_gmt":"2026-02-03T09:42:28","guid":{"rendered":"https:\/\/jobuzo.com\/en\/compact-ai-model-from-chinas-stepfun-outshines-rivals-from-deepseek-moonshot\/"},"modified":"2026-02-03T09:42:28","modified_gmt":"2026-02-03T09:42:28","slug":"compact-ai-model-from-chinas-stepfun-outshines-rivals-from-deepseek-moonshot","status":"publish","type":"post","link":"https:\/\/jobuzo.com\/en\/compact-ai-model-from-chinas-stepfun-outshines-rivals-from-deepseek-moonshot\/","title":{"rendered":"Compact AI model from China\u2019s StepFun outshines rivals from DeepSeek, Moonshot"},"content":{"rendered":"<div>\n<div datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1xdhyk6 ec74h0k0\" readability=\"8.0827067669173\">Chinese artificial intelligence start-up StepFun has unveiled a lightweight AI model that it says punches above its weight, rivalling larger systems from domestic competitors including DeepSeek and Moonshot AI as c<span data-qa=\"Component-Text\" class=\"css-0 ef9u0v00\">ompetition intensifies in the country&rsquo;s AI sector<\/span>.<\/div>\n<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">The Shanghai-based AI lab said on Monday its latest Step 3.5 Flash model was designed to deliver advanced reasoning and agentic capabilities while maintaining efficiency.<\/p>\n<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">Despite its relatively modest size of about 196 billion parameters &ndash; far smaller than Moonshot AI&rsquo;s Kimi K2.5 with 1 trillion parameters or DeepSeek V3.2 with 671 billion parameters &ndash; Step 3.5 Flash outperformed its larger rivals across several benchmark tests measuring agentic, reasoning and coding capabilities, according to the company&rsquo;s self-reported results.<\/p>\n<div data-qa=\"InlineAdSlot-Container\" class=\"css-zl1inp e11v3ui14\">\n<div class=\"e11v3ui10 e11v3ui13 css-1evd7i0 e1flwkbl0\" data-qa=\"AdSlot-Container\">\n<p>Advertisement<\/p>\n<\/div>\n<\/div>\n<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">Parameters are the variables that encode an AI system&rsquo;s &ldquo;intelligence&rdquo;, with a larger number usually indicating stronger performance.<\/p>\n<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">Step 3.5 Flash topped four reasoning benchmarks, including AIME 2025 and IMOAnswerBench, outperforming leading systems from DeepSeek, Moonshot AI, Zhipu AI and MiniMax, and trailing only Microsoft-backed OpenAI in certain tests.<\/p>\n<div class=\"image-inline-container e1a5rv550 css-1llrc1m e1yqhwb40\" data-qa=\"Component-renderMap-StyledDiv\">\n<div class=\"image-inline caption e1fvabeq0 css-19sk4h4 ea9pn0s0\" data-qa=\"Component-Container\">\n<figure class=\"image-inline caption ea9pn0s1 css-1qeofuq e1gf69pb0\" data-qa=\"ArticleImage-ArticleImageContainer\">\n<div data-qa=\"ArticleImage-handleRenderImage-ImageContainer\" class=\"css-0 e1gf69pb3\"><\/div><figcaption data-qa=\"ArticleImage-DescriptionContainer\" class=\"css-1ixmelf e1gf69pb1\">StepFun says its lightweight model was designed to deliver advanced reasoning and agentic capabilities while maintaining efficiency. Photo: Handout<\/figcaption><\/figure>\n<\/div>\n<\/div>\n<div datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1xdhyk6 ec74h0k0\" readability=\"14.189189189189\">The model&rsquo;s compact size and <span data-qa=\"Component-Text\" class=\"css-0 ef9u0v00\">focus on reasoning<\/span> were deliberate choices, according to Zhu Yibo, StepFun&rsquo;s co-founder and chief technology officer. Zhu said the team had prioritised &ldquo;strong logic capability, efficient context window and fast speed&rdquo; when developing the new system, which was purpose-built for the AI agent era.<\/div>\n<\/div>\n<p><sub>Compact AI model from China&rsquo;s StepFun outshines rivals from DeepSeek, Moonshot<\/sub><\/p>\n","protected":false},"excerpt":{"rendered":"<p>Chinese artificial intelligence start-up StepFun has unveiled a lightweight AI model that it says punches above its weight, rivalling larger systems from domestic competitors including DeepSeek and Moonshot AI as competition intensifies in the country&rsquo;s AI sector. The Shanghai-based AI lab said on Monday its latest Step 3.5 Flash model was designed to deliver advanced&#8230;<\/p>\n<p class=\"more-link-wrap\"><a href=\"https:\/\/jobuzo.com\/en\/compact-ai-model-from-chinas-stepfun-outshines-rivals-from-deepseek-moonshot\/\" class=\"more-link\">Read More<span class=\"screen-reader-text\"> &ldquo;Compact AI model from China\u2019s StepFun outshines rivals from DeepSeek, Moonshot&rdquo;<\/span> &raquo;<\/a><\/p>\n","protected":false},"author":1,"featured_media":14998,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-14997","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news"],"_links":{"self":[{"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/posts\/14997","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/comments?post=14997"}],"version-history":[{"count":0,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/posts\/14997\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/media\/14998"}],"wp:attachment":[{"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/media?parent=14997"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/categories?post=14997"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/jobuzo.com\/en\/wp-json\/wp\/v2\/tags?post=14997"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}