{"id":45799,"date":"2025-11-09T14:35:47","date_gmt":"2025-11-09T06:35:47","guid":{"rendered":"https:\/\/www.1ai.net\/?p=45799"},"modified":"2025-11-09T14:35:47","modified_gmt":"2025-11-09T06:35:47","slug":"%e6%9c%88%e4%b9%8b%e6%9a%97%e9%9d%a2-kimi-k2-thinking-%e8%ae%ad%e7%bb%83%e6%88%90%e6%9c%ac%e8%a2%ab%e6%9b%9d%e4%bb%85-460-%e4%b8%87%e7%be%8e%e5%85%83%ef%bc%8c%e6%80%a7%e8%83%bd%e8%b6%85%e8%b6%8a","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/45799.html","title":{"rendered":"The dark side of the moon Kimi K2 Tinking training costs were exposed to $4.6 million and performance exceeded the OpenAI GPT model of billions of dollars invested"},"content":{"rendered":"<p>On November 9th<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e6%9c%88%e4%b9%8b%e6%9a%97%e9%9d%a2\" title=\"[Sees articles with labels]\" target=\"_blank\" >Dark Side of the Moon<\/a>(Moonshot AI)<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%bc%80%e6%ba%90\" title=\"[View articles tagged with [open source]]\" target=\"_blank\" >Open Source<\/a><a href=\"https:\/\/www.1ai.net\/en\/tag\/%e6%80%9d%e8%80%83%e6%a8%a1%e5%9e%8b\" title=\"[Sees articles with [thinking model] labels]\" target=\"_blank\" >Thinking Model<\/a> \u2014\u2014<a href=\"https:\/\/www.1ai.net\/en\/tag\/kimi\" title=\"[View articles tagged with [Kimi]]\" target=\"_blank\" >Kimi<\/a> K2 Thinking\u3002<\/p>\n<p>According to the dark side of the month, Kimi K2 Thinking achieved an excellent performance of 44.9% in the final human examination (HLE), which went beyond advanced models such as GPT-5, Grok-4, Claude 4.5\u3002<\/p>\n<p>However, CNBC quoted sources as saying that Kimi K2 Thinking's training cost was only $4.6 million (note: the current exchange rate is about RMB 327.7 million), lower than the $5.6 million of DeepSeek V3 - in contrast to the fact that the US$ 5.6 million was spent on training for the Kimi K2 Thinking model<a href=\"https:\/\/www.1ai.net\/en\/tag\/openai\" title=\"[View articles tagged with [OpenAI]]\" target=\"_blank\" >OpenAI<\/a> Investment in related research and development amounts to billions of dollars\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-45800\" title=\"a8359e7j00t5g4y7001hd000ugwm\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2025\/11\/a83509e7j00t5g4y7001hd000u000gwm.jpg\" alt=\"a8359e7j00t5g4y7001hd000ugwm\" width=\"1080\" height=\"608\" \/><\/p>\n<p>Unlike many closed-source models in the market, Kimi K2 Thinking adopted open-source strategies and issued model weights, training scripts, data matching and assessment tool chains and allowed commercial use. This initiative has significantly reduced the threshold for use by developers and contributed to the spread of AI technology\u3002<\/p>","protected":false},"excerpt":{"rendered":"<p>On November 9, Moonshot AI launched its most powerful open-source thinking model so far on Thursday. According to the dark side of the month, Kimi K2 Thinking achieved an excellent performance of 44.9% in the final human examination (HLE), which went beyond advanced models such as GPT-5, Grok-4, Claude 4.5. However, CNBC quotes sources that Kimi K2 Thinking training costs only $4.6 million (note: current exchange rate is about RMB 327.7 million), lower than the 560 of DeepSeek V3<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[1814,190,219,7851,1168],"collection":[],"class_list":["post-45799","post","type-post","status-publish","format-standard","hentry","category-news","tag-kimi","tag-openai","tag-219","tag-7851","tag-1168"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/45799","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=45799"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/45799\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=45799"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=45799"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=45799"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=45799"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}