{"id":59328,"date":"2026-10-24T11:15:05","date_gmt":"2026-10-24T03:15:05","guid":{"rendered":"https:\/\/www.1ai.net\/?p=59328"},"modified":"2026-10-24T11:15:05","modified_gmt":"2026-10-24T03:15:05","slug":"kimi-k3-1-pro-%e6%ad%a3%e5%bc%8f%e4%b8%8a%e7%ba%bf%ef%bc%8c%e6%94%af%e6%8c%81%e7%99%be%e4%b8%87%e4%b8%8a%e4%b8%8b%e6%96%87%ef%bc%8c%e5%ae%8c%e6%95%b4%e6%9d%83%e9%87%8d%e5%b0%86%e5%9c%a8-11-%e6%9c%88-4","status":"publish","type":"post","link":"http:\/\/www.1ai.net\/en\/59328.html","title":{"rendered":"Kimi K3.1 Pro is officially online to support millions of context, full weight will be released by 4 November"},"content":{"rendered":"<p>October 24th.<a href=\"http:\/\/www.1ai.net\/en\/tag\/%e6%9c%88%e4%b9%8b%e6%9a%97%e9%9d%a2\" title=\"[Sees articles with labels]\" target=\"_blank\" rel=\"noopener\">Dark Side of the Moon<\/a>Official Release <a href=\"http:\/\/www.1ai.net\/en\/tag\/kimi\" title=\"[View articles tagged with [Kimi]]\" target=\"_blank\" rel=\"noopener\">Kimi<\/a> K3.1 Pro and the simultaneous launch of the Highspeed high-speed version to support the 1 million token context. The official plan is to publish the full model weights by November 4\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-59329\" title=\"011d9e77j00tne6bf0051d000ukm\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/10\/011d9e77j00tne6bf0051d000u000ckm.jpg\" alt=\"011d9e77j00tne6bf0051d000ukm\" width=\"1080\" height=\"452\" \/><\/p>\n<p>K3.1 Pro is based on Kimi K3 to increase the calculus of training and to jointly enhance learning in hundreds of downstream missions. The output speed of the Highspeed version is said to be 2 to 3 times the standard version\u3002<\/p>\n<p>In the assessment published on the dark side of the month, the Kerminal-Bench 4.0 of K3.1 Pro was divided into 65.7%; in the internal factual accuracy pressure test, the hallucinogenic rate of the individual statements was reduced from 23.4% to 8.3%\u3002<\/p>\n<p>According to Artificial Analysis, K3.1 Pro scored 55 on its intelligence index, ranked 4, 11 points above K3, second only to Claude Opus 5.5, Claude Sonnet 5.5 and Muse Wonder Full\u3002<\/p>\n<p>The model ' s input, output and CAL prices were $33% lower than K3 at $2, $10 and $0.20 per million token, respectively\u3002<\/p>\n<p>The company also stated that, in order to reduce the risk of cyber-security abuse after open weights, relevant data had been removed from medium-term training and enhanced learning environments, while retaining generic programming tasks. The Kimi enterprise version will not be enabled by default when it is on line and will need to be opened by the organizer\u3002<\/p>","protected":false},"excerpt":{"rendered":"<p>On October 24, Kimi K3.1 Pro was officially released on the dark side of the month, and the Highspeed high-speed version was launched simultaneously to support the 1 million token context. The official plan is to publish the full model weights by November 4. K3.1 Pro is based on Kimi K3 to increase the calculus of training and to jointly enhance learning in hundreds of downstream missions. The output speed of the Highspeed version is said to be 2 to 3 times the standard version. In the assessment published on the dark side of the moon, Kerminal-Bench 4.0 of K3.1 Pro is divided into 65.7%; in the internal factual accuracy pressure test, the hallucinogenic rate of the individual statement is K3<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[146],"tags":[1814,1168],"collection":[],"class_list":["post-59328","post","type-post","status-publish","format-standard","hentry","category-news","tag-kimi","tag-1168"],"acf":[],"_links":{"self":[{"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/59328","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=59328"}],"version-history":[{"count":0,"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/59328\/revisions"}],"wp:attachment":[{"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=59328"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=59328"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=59328"},{"taxonomy":"collection","embeddable":true,"href":"http:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=59328"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}