{"id":1313,"date":"2025-07-14T11:47:33","date_gmt":"2025-07-14T11:47:33","guid":{"rendered":"https:\/\/finzexpert.com\/blog\/alibaba-backed-moonshot-releases-kimi-k2-ai-rivaling-chatgpt-claude\/"},"modified":"2025-07-14T11:47:33","modified_gmt":"2025-07-14T11:47:33","slug":"alibaba-backed-moonshot-releases-kimi-k2-ai-rivaling-chatgpt-claude","status":"publish","type":"post","link":"https:\/\/finzexpert.com\/blog\/alibaba-backed-moonshot-releases-kimi-k2-ai-rivaling-chatgpt-claude\/","title":{"rendered":"Alibaba-backed Moonshot releases Kimi K2 AI rivaling ChatGPT, Claude"},"content":{"rendered":"\n<div id=\"RegularArticle-ArticleBody-5\" data-module=\"ArticleBody\" data-test=\"articleBody-2\" data-analytics=\"RegularArticle-articleBody-5-2\"><span class=\"HighlightShare-hidden\" style=\"top:0;left:0\"\/><\/p>\n<div class=\"InlineImage-imageEmbed\" id=\"ArticleBody-InlineImage-108171234\" data-test=\"InlineImage\">\n<div class=\"InlineImage-wrapper\">\n<div>\n<p>An AI sign at the MWC Shanghai tech show on June 19, 2025.<\/p>\n<p>Bloomberg | Bloomberg | Getty Images<\/p>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"group\">\n<p>BEIJING \u2014 The latest Chinese generative artificial intelligence model to take on OpenAI&#8217;s ChatGPT is offering coding capabilities \u2014 at a lower price.<\/p>\n<p>Alibaba-backed startup Moonshot <a href=\"https:\/\/x.com\/Kimi_Moonshot\/status\/1943687594560332025\" target=\"_blank\">released on late Friday night <\/a>its Kimi K2 model: a low-cost, open source large language model \u2014 the two factors that underpinned China-based DeepSeek&#8217;s industry disruption in January. Open-source technology provides source code access for free, an approach that few U.S. tech giants have taken, other than Meta and Google to some extent.<\/p>\n<p>Coincidentally, OpenAI CEO Sam Altman <a href=\"https:\/\/x.com\/sama\/status\/1943837550369812814\" target=\"_blank\">announced early Saturday<\/a> that there would be an indefinite delay of its first open-source model yet again due to safety concerns. OpenAI did not immediately respond to a CNBC request for comment on Kimi K2.<\/p>\n<\/div>\n<div role=\"region\" aria-labelledby=\"Placeholder-ArticleBody-Video-108170982\">\n<div role=\"button\" tabindex=\"0\" id=\"Placeholder-ArticleBody-Video-108170982\" class=\"PlaceHolder-wrapper\" data-vilynx-id=\"7000382123\" data-test=\"VideoPlaceHolder\">\n<div class=\"InlineVideo-videoEmbed\" id=\"InlineVideo-0\" data-test=\"InlineVideo\">\n<div class=\"InlineVideo-wrapper\">\n<div class=\"InlineVideo-inlineThumbnailContainer\"><span class=\"InlineVideo-videoButton\"\/><span\/><\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n<div class=\"group\">\n<p>One of Kimi K2&#8217;s strengths is in writing computer code for applications, an area in which businesses see potential to reduce or replace staff with generative AI. OpenAI&#8217;s U.S. rival Anthropic <a href=\"https:\/\/www.anthropic.com\/news\/claude-4\" target=\"_blank\" rel=\"noopener\">focused on coding with its Claude Opus 4 model<\/a> released in late May.<\/p>\n<p>In its release announcement on social media platforms X and GitHub, Moonshot claimed Kimi K2 surpassed Claude Opus 4 on two benchmarks, and had better overall performance than OpenAI&#8217;s coding-focused GPT-4.1 model, based on several industry metrics.<\/p>\n<p>&#8220;No doubt [Kimi K2 is] a globally competitive model, and it&#8217;s open sourced,&#8221; Wei Sun, principal analyst in artificial intelligence at Counterpoint, said in an email Monday.<\/p>\n<\/div>\n<h2 class=\"ArticleBody-subtitle\"><a id=\"headline0\"\/>Cheaper option<\/h2>\n<div class=\"group\">\n<p>&#8220;On top of that, it has lower token costs, making it attractive for large-scale or budget-sensitive deployments,&#8221; she said.<\/p>\n<p>The new K2 model is available via Kimi&#8217;s app and browser interface for free unlike ChatGPT or Claude, which charge monthly subscriptions for their latest AI models.<\/p>\n<p>Kimi is also only charging 15 cents for every 1 million input tokens, and $2.50 per 1 million output tokens, according to its website. Tokens are a way of <a href=\"https:\/\/help.openai.com\/en\/articles\/4936856-what-are-tokens-and-how-to-count-them\" target=\"_blank\" rel=\"noopener\">measuring data for AI model processing<\/a>.<\/p>\n<p>In contrast, Claude Opus 4 <a href=\"https:\/\/www.anthropic.com\/claude\/opus\" target=\"_blank\" rel=\"noopener\">charges 100 times more<\/a> for input \u2014 $15 per million tokens \u2014 and 30 times more for output \u2014 $75 per million tokens. Meanwhile, for every one million tokens, <a href=\"https:\/\/openai.com\/index\/gpt-4-1\/\" target=\"_blank\" rel=\"noopener\">GPT-4.1 charges<\/a> $2 for input and $8 for output.<\/p>\n<p>Moonshot AI <a href=\"https:\/\/github.com\/MoonshotAI\/Kimi-K2?tab=License-1-ov-file#readme\" target=\"_blank\" rel=\"noopener\">said on GitHub<\/a> that developers can use K2 however they wish, with the only requirement that they display &#8220;Kimi K2&#8221; on the user interface if the commercial product or service has more than 100 million monthly active users, or makes the equivalent of $20 million in monthly revenue.<\/p>\n<\/div>\n<h2 class=\"ArticleBody-subtitle\"><a id=\"headline1\"\/>Hot AI market<\/h2>\n<div class=\"group\">\n<p>Initial reviews of K2 on both English and Chinese social media have largely been positive, although there are <a href=\"https:\/\/zhuanlan.zhihu.com\/p\/1927323824300496714\" target=\"_blank\" rel=\"noopener\">some reports of hallucinations<\/a>, a prevalent issue in generative AI, in which the models make up information.<\/p>\n<p>Still, K2 is &#8220;the first model I feel comfortable using in production since Claude 3.5 Sonnet,&#8221; Pietro Schirano, <a href=\"https:\/\/x.com\/skirano\/status\/1944475540951621890?s=46\" target=\"_blank\">founder of startup MagicPath<\/a> that offers AI tools for design, said in a post on X.<\/p>\n<p>Moonshot has open sourced some of its prior AI models. The company&#8217;s chatbot surged in popularity early last year as China&#8217;s alternative to ChatGPT, which isn&#8217;t officially available in the country. But similar chatbots from ByteDance and Tencent have since crowded the market, while tech giant Baidu has revamped its core search engine with AI tools.<\/p>\n<p>Kimi&#8217;s latest AI release comes as investors eye Chinese alternatives to U.S. tech in the global AI competition. <\/p>\n<p>Still, despite the excitement about DeepSeek, the privately-held company has yet to announce a major upgrade to its R1 and V3 model. Meanwhile, Manus AI, a Chinese startup that emerged earlier this year as another DeepSeek-type upstart, has<a href=\"https:\/\/manus.im\/about\" target=\"_blank\" rel=\"noopener\"> relocated its headquarters to Singapore<\/a>.<\/p>\n<p>Over in the U.S., OpenAI also has yet to reveal GPT-5.<\/p>\n<p>Work on GPT-5 may be taking up engineering resources, preventing OpenAI from progressing on its open-source model, Counterpoint&#8217;s Sun said, adding that it&#8217;s challenging to release a powerful open-source model without undermining the competitive advantage of a proprietary model.<\/p>\n<\/div>\n<h2 class=\"ArticleBody-subtitle\"><a id=\"headline2\"\/>Grok 4 competitor<\/h2>\n<div class=\"group\">\n<p>Kimi K2 is not the company&#8217;s only recent release. Moonshot launched a Kimi research model last month and claimed it matched Google&#8217;s Gemini Deep Research &#8216;s 26.9 score and <a href=\"https:\/\/x.com\/Kimi_Moonshot\/status\/1936105761358368944\" target=\"_blank\">beat OpenAI&#8217;s version on a benchmark called &#8220;Humanity&#8217;s Last Exam.&#8221;<\/a><\/p>\n<p>The Kimi research model even got a mention last week during Elon Musk&#8217;s <a href=\"https:\/\/x.com\/xai\/status\/1943158495588815072?s=46\" target=\"_blank\">xAI release of Grok 4<\/a> \u2014 which scored 25.4 on its own on the &#8220;Humanity&#8217;s Last Exam&#8221; benchmark, but <a href=\"https:\/\/x.com\/xai\/status\/1943786241306374426\/photo\/1\" target=\"_blank\">attained a 44.4 score when allowed to use a variety of AI tools and web search.<\/a><\/p>\n<p>&#8220;Kimi-Researcher represents a paradigm shift in agentic AI,&#8221; said Winston Ma, adjunct professor at NYU School of Law. He was referring to AI&#8217;s capability of simultaneously making several decisions on its own to complete a complex task.<\/p>\n<p>&#8220;Instead of merely generating fluent responses, it demonstrates autonomous reasoning at an expert level \u2014 the kind of complex cognitive work previously missing from LLMs,&#8221; Ma said. He is also author of &#8220;The Digital War: How China&#8217;s Tech Power Shapes the Future of AI, Blockchain and Cyberspace.&#8221;<\/p>\n<p><em>\u2014 CNBC&#8217;s Victoria Yeo contributed to this report.<\/em><\/p>\n<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>An AI sign at the MWC Shanghai tech show on June 19, 2025. Bloomberg | Bloomberg | Getty Images BEIJING \u2014 The latest Chinese generative artificial intelligence model to take on OpenAI&#8217;s ChatGPT is offering coding capabilities \u2014 at a lower price. Alibaba-backed startup Moonshot released on late Friday night its Kimi K2 model: a [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":1314,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8],"tags":[1101,1105,1106,1103,1102,695,1104],"class_list":["post-1313","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-loan","tag-alibababacked","tag-chatgpt","tag-claude","tag-kimi","tag-moonshot","tag-releases","tag-rivaling"],"_links":{"self":[{"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/posts\/1313","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/comments?post=1313"}],"version-history":[{"count":0,"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/posts\/1313\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/media\/1314"}],"wp:attachment":[{"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/media?parent=1313"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/categories?post=1313"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/finzexpert.com\/blog\/wp-json\/wp\/v2\/tags?post=1313"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}