{"id":40706,"date":"2025-08-03T13:34:01","date_gmt":"2025-08-03T05:34:01","guid":{"rendered":"https:\/\/www.1ai.net\/?p=40706"},"modified":"2025-08-03T13:34:01","modified_gmt":"2025-08-03T05:34:01","slug":"%e6%98%86%e4%bb%91%e4%b8%87%e7%bb%b4%e5%8f%91%e5%b8%83%e5%b9%b6%e5%bc%80%e6%ba%90%e5%85%a8%e6%96%b0%e6%8e%a8%e7%90%86%e5%a4%a7%e6%a8%a1%e5%9e%8b-mindlink%ef%bc%8c%e5%9f%ba%e4%ba%8e-qwen3-32b-%e5%92%8c","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/40706.html","title":{"rendered":"KunlunWanwei Releases and Open Sources New Reasoning Big Model MindLink, Based on Qwen3-32B and Qwen2.5-72B for Post-Training"},"content":{"rendered":"<p>August 3 News.<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e6%98%86%e4%bb%91%e4%b8%87%e7%bb%b4\" title=\"[Sees articles with [Konlen] tags]\" target=\"_blank\" >Kunlun Wanwei<\/a>Announcing the launch and<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>all-new<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e6%8e%a8%e7%90%86%e5%a4%a7%e6%a8%a1%e5%9e%8b\" title=\"Look at the article that contains the labels\" target=\"_blank\" >Large model of inference<\/a> Skywork <a href=\"https:\/\/www.1ai.net\/en\/tag\/mindlink\" title=\"[See articles containing [MindLink] labels]\" target=\"_blank\" >MindLink<\/a>.<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-40707\" title=\"c4238c54j00t0ekrj005rd000v900nip\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2025\/08\/c4238c54j00t0ekrj005rd000v900nip.jpg\" alt=\"c4238c54j00t0ekrj005rd000v900nip\" width=\"1125\" height=\"846\" \/><\/p>\n<p>It is reported that this new reasoning paradigm is able to adaptively integrate reasoned and non-reasoned generated responses according to the task difficulty, which greatly reduces the reasoning cost and enhances the readability and effectiveness of the reasoning process.<\/p>\n<p>The data show that the Skywork MindLink model has been the first in 10 large model evaluations with no external tool configuration, in the final human examination (Humanity's Last Exam), in four first evaluations in complex mathematics (Ossay) (USAMO 2025, AIME 2024, AIMS 2025, HMMT 2025), and also in the field of general knowledge GPQA-diamond and MMLU-pro\u3002<\/p>\n<p>Core capabilities of the Skywork MindLink model include:<\/p>\n<ul>\n<li>The new reasoning paradigm Plan-based Reasoning removes the \"think\" label, reduces the cost of reasoning, and improves the ability to have multi-round conversations.<\/li>\n<li>A new mathematical approach is presented that can attempt to analyze the effectiveness of CoT and Plan-based Reasoning.<\/li>\n<li>Proposes a set of adaptive reasoning mechanisms capable of integrating reasoned and non-reasoned generated responses based on task difficulty.<\/li>\n<li>Post-training based on Qwen3-32B and Qwen2.5-72B improves the original model capability while saving a lot of training cost, and more size models will be open-sourced in the future.<\/li>\n<\/ul>\n<p>1AI noted that currently KunlunWanwei already has big models such as Tiangong Language Big Model, Tiangong Multimodal Big Model, SWE Code Big Model, Agent Big Model, Video Big Model, 3D Big Model, Music Big Model, Audio Big Model, and so on.<\/p>\n<p>Official Address:<\/p>\n<ul>\n<li><strong>72B Model weights<\/strong>: MindLink 72B-0801: https:\/\/huggingface.co\/Skywork\/MindLink-72B-0801<\/li>\n<li><strong>Technical Report<\/strong>:: https:\/\/github.com\/SkyworkAI\/MindLink\/blob\/main\/mindlink.pdf<\/li>\n<li><strong>code repository<\/strong>:: https:\/\/github.com\/SkyworkAI\/MindLink<\/li>\n<\/ul>","protected":false},"excerpt":{"rendered":"<p>On August 3rd, Kuala Lumpur announced the launch and launch of a new big model of reasoning Skywork MindLink. It was stated that this whole new paradigm of reasoning, which allows for the adaptation of integration and non-adjective responses to the difficulties of the task, significantly reduces the cost of reasoning and enhances the readability and effectiveness of the reasoning process. According to the data, the Skywork MindLink model was the first in 10 large model evaluations with no external tool configuration, in the final human examination (Humanity's Last Exam) and in the complex field of mathematics (Osaille), in the first of 4 evaluations (USAMO 2025, AIME 2024, AIME 2025, AIME 2025, AME 2025, AME 2025, AME 205, AME 205, AMS 2025, AMS 2025, AMS 2025, AMS 2025, AMS 2025<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[7332,219,6260,1050],"collection":[],"class_list":["post-40706","post","type-post","status-publish","format-standard","hentry","category-news","tag-mindlink","tag-219","tag-6260","tag-1050"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/40706","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=40706"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/40706\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=40706"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=40706"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=40706"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=40706"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}