{"id":58549,"date":"2026-09-22T09:35:42","date_gmt":"2026-09-22T01:35:42","guid":{"rendered":"https:\/\/www.1ai.net\/?p=58549"},"modified":"2026-09-22T09:35:42","modified_gmt":"2026-09-22T01:35:42","slug":"%e5%b0%8f%e7%b1%b3-mimo-v2-6-%e5%bc%80%e6%ba%90%ef%bc%9a%e6%94%af%e6%8c%81-1m-%e4%b8%8a%e4%b8%8b%e6%96%87%e8%ae%ad%e7%bb%83%ef%bc%8cagent-%e8%83%bd%e5%8a%9b%e5%a4%a7%e5%b9%85%e5%8d%87%e7%ba%a7","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/58549.html","title":{"rendered":"MiMo-V2.6 Open Source: Support 1M context training, Agent capacity significantly upgraded"},"content":{"rendered":"<p>The news of September 22nd<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%b0%8f%e7%b1%b3\" title=\"[View articles tagged with [Xiaomi]]\" target=\"_blank\" >Millet<\/a>Published yesterday<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> Xiaoomi MiMo-V2.6 series with Pro and Flash originals<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%85%a8%e6%a8%a1%e6%80%81%e6%a8%a1%e5%9e%8b\" title=\"[Sees articles with labels]\" target=\"_blank\" >holomodal model<\/a>.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-58550\" title=\"8342a048jlqsej003d000u000e2m\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/09\/8342a048j00tlqsej003dd000u000e2m.jpg\" alt=\"8342a048jlqsej003d000u000e2m\" width=\"1080\" height=\"506\" \/><\/p>\n<p>MiMo-V2.6-Pro received 46 points in Artificial Analysis Inteligence Index, exceeding Kimi K3 and Qwen3.8 Max, but still gaps with closed-source models such as Claude Fable 5.1, GPT-6 Astra\u3002<\/p>\n<p>It is known that the MiMo-V2.6-Flash and Pro Live RL training lasted less than six days, at a cost of approximately $850,000 and $2.62 million, respectively, and completed 30 steps each, each resulting in a cumulative trajectory of about 750,000\u3002<\/p>\n<p>After the training, the two models were raised by approximately 17 and 14 points on DeepSWE v1.1, respectively. The training was updated on a single basis with 1568 samples, supporting 1M context length, and one-step training token reached 2.7 billion to 3.7 billion\u3002<\/p>\n<p>In terms of capability, MiMo-V2.6 further integrates multimodular perception, 3D spatial reasoning and the Computer Use capability, which can be used to generate 3D Open World games, Blender modelling, control simulation mechanical arms, and to operate common office and productivity software\u3002<\/p>\n<p>Mimo Desktop desktop client with Mimo-V2.6-Pro providing a UltraSpeed mode with a maximum output speed of 20 times\u3002<\/p>\n<p>The MiMo-V2.6 series has also landed on the MiMo Open Platform, and API pricing follows the previous generation: Pro general input and output prices are $1 and $2 per million Token 3 and 6\u3002<\/p>","protected":false},"excerpt":{"rendered":"<p>On September 22nd, Mi was officially released yesterday and opened with the Xiaomi MiMo-V2.6 series, which contains both Pro and Flash original whole-state models. MiMo-V2.6-Pro received 46 points in Artificial Analysis Inteligence Index, exceeding Kimi K3 and Qwen3.8 Max, but still gaps with closed-source models such as Claude Fable 5.1, GPT-6 Astra. It is known that MiMo-V2.6-Flash has less than six days of Live RL training and that training costs are around 8<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[5569,1114,219],"collection":[],"class_list":["post-58549","post","type-post","status-publish","format-standard","hentry","category-news","tag-5569","tag-1114","tag-219"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/58549","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=58549"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/58549\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=58549"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=58549"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=58549"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=58549"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}