{"id":42939,"date":"2025-09-12T11:50:29","date_gmt":"2025-09-12T03:50:29","guid":{"rendered":"https:\/\/www.1ai.net\/?p=42939"},"modified":"2025-09-12T11:50:29","modified_gmt":"2025-09-12T03:50:29","slug":"%e9%98%bf%e9%87%8c%e4%ba%91%e5%8f%91%e5%b8%83%e9%80%9a%e4%b9%89-qwen3-next-%e5%9f%ba%e7%a1%80%e6%a8%a1%e5%9e%8b%e6%9e%b6%e6%9e%84%ef%bc%8c%e5%bc%80%e6%ba%9080b-a3b%e7%b3%bb%e5%88%97%e6%a8%a1%e5%9e%8b","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/42939.html","title":{"rendered":"Aliyun release generic Qwen3-Next basic model structure, open source 80B-A3B series model"},"content":{"rendered":"<p>\u4eca\u5929\u51cc\u6668\uff0c\u963f\u91cc\u901a\u4e49\u53d1\u5e03\u4e86\u4e0b\u4e00\u4ee3<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%9f%ba%e7%a1%80%e6%a8%a1%e5%9e%8b%e6%9e%b6%e6%9e%84\" title=\"\u3010\u67e5\u770b\u542b\u6709[\u57fa\u7840\u6a21\u578b\u67b6\u6784]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">\u57fa\u7840\u6a21\u578b\u67b6\u6784<\/a> Qwen3-Next\uff0c\u5e76<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%bc%80%e6%ba%90\" title=\"\u3010\u67e5\u770b\u542b\u6709[\u5f00\u6e90]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">\u5f00\u6e90<\/a>\u4e86\u57fa\u4e8e\u8be5\u67b6\u6784\u7684 Qwen3-Next-80B-A3B \u7cfb\u5217\u6a21\u578b\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-42940\" title=\"62ec15bej00t2gil0002pd000u000gwm\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2025\/09\/62ec15bej00t2gil0002pd000u000gwm.jpg\" alt=\"62ec15bej00t2gil0002pd000u000gwm\" width=\"1080\" height=\"608\" \/><\/p>\n<p>\u5b98\u65b9\u8868\u793a\uff0c\u5176\u8ba4\u4e3a Context Length Scaling \u548c Total Parameter Scaling \u662f\u672a\u6765\u5927\u6a21\u578b\u53d1\u5c55\u7684\u4e24\u5927\u8d8b\u52bf\u3002\u4e3a\u4e86\u8fdb\u4e00\u6b65\u63d0\u5347\u6a21\u578b\u5728\u957f\u4e0a\u4e0b\u6587\u548c\u5927\u89c4\u6a21\u603b\u53c2\u6570\u4e0b\u7684\u8bad\u7ec3\u548c\u63a8\u7406\u6548\u7387\uff0c\u901a\u4e49\u5343\u95ee\u8bbe\u8ba1\u4e86\u5168\u65b0\u7684 Qwen3-Next \u7684\u6a21\u578b\u7ed3\u6784\u3002<\/p>\n<p>\u636e\u4ecb\u7ecd\uff0cQwen3-Next \u76f8\u6bd4 Qwen3 \u7684 MoE \u6a21\u578b\u7ed3\u6784,\u8fdb\u884c\u4e86\u4ee5\u4e0b\u6838\u5fc3\u6539\u8fdb\uff1a\u6df7\u5408\u6ce8\u610f\u529b\u673a\u5236\u3001\u9ad8\u7a00\u758f\u5ea6 MoE \u7ed3\u6784\u3001\u4e00\u7cfb\u5217\u8bad\u7ec3\u7a33\u5b9a\u53cb\u597d\u7684\u4f18\u5316\uff0c\u4ee5\u53ca\u63d0\u5347\u63a8\u7406\u6548\u7387\u7684\u591a token \u9884\u6d4b\u673a\u5236\u3002<\/p>\n<p>\u57fa\u4e8e Qwen3-Next \u7684\u6a21\u578b\u7ed3\u6784\uff0c\u901a\u4e49\u8fd8\u8bad\u7ec3\u4e86 Qwen3-Next-80B-A3B-Base \u6a21\u578b\u3002\u8be5\u6a21\u578b\u62e5\u6709 800 \u4ebf\u53c2\u6570\u4ec5\u6fc0\u6d3b 30 \u4ebf\u53c2\u6570\uff0c\u5e76\u5b9e\u73b0\u4e86\u4e0eQwen3-32B dense \u6a21\u578b\u76f8\u8fd1\u751a\u81f3\u7565\u597d\u7684\u6027\u80fd\u3002<\/p>\n<p>\u800c Qwen3-Next-80B-A3B-Base \u8bad\u7ec3\u6210\u672c\uff08GPU hours\uff09\u4ec5\u4e3a Qwen3-32B \u7684\u5341\u5206\u4e4b\u4e00\u4e0d\u5230\uff0c\u5728 32k \u4ee5\u4e0a\u7684\u4e0a\u4e0b\u6587\u4e0b\u7684\u63a8\u7406\u541e\u5410\u5219\u662f Qwen3-32B \u7684\u5341\u500d\u4ee5\u4e0a\uff0c\u5b9e\u73b0\u4e86\u6781\u81f4\u7684\u8bad\u7ec3\u548c\u63a8\u7406\u6027\u4ef7\u6bd4\u3002<\/p>\n<p>\u540c\u65f6\uff0c\u901a\u4e49\u8fd8\u57fa\u4e8e Qwen3-Next-80B-A3B-Base \u578b, \u540c\u6b65\u5f00\u53d1\u5e76\u53d1\u5e03\u4e86 \u00a0Qwen3-Next-80B-A3B-Instruct\u00a0\u4e0e\u00a0<strong>Qwen3-Next-80B-A3B-Thinking\uff1a<\/strong><\/p>\n<p>Qwen3-Next-80B-A3B-Instruct \u4e0e\u65d7\u8230\u6a21\u578b Qwen3-235B-A22B-Instruct-2507 \u8868\u73b0\u76f8\u5f53\uff0c\u540c\u65f6\u5728 256K \u8d85\u957f\u4e0a\u4e0b\u6587\u5904\u7406\u4efb\u52a1\u4e2d\u5c55\u73b0\u51fa\u663e\u8457\u4f18\u52bf\uff1b<\/p>\n<p>Qwen3-Next-80B-A3B-Thinking \u5728\u590d\u6742\u63a8\u7406\u4efb\u52a1\u4e0a\u8868\u73b0\u5353\u8d8a\uff0c\u4e0d\u4ec5\u4f18\u4e8e\u9884\u8bad\u7ec3\u6210\u672c\u66f4\u9ad8\u7684 Qwen3-30B-A3B-Thinking-2507 \u4e0e Qwen3-32B-Thinking\uff0c\u66f4\u5728\u591a\u9879\u57fa\u51c6\u6d4b\u8bd5\u4e2d\u8d85\u8d8a\u95ed\u6e90\u6a21\u578b Gemini-2.5-Flash-Thinking\uff0c\u90e8\u5206\u5173\u952e\u6307\u6807\u5df2\u903c\u8fd1 Qwen3-235B-A22B-Thinking-2507\u3002<\/p>\n<p>\u76ee\u524d\uff0c\u65b0\u6a21\u578b\u5df2\u4e0a\u7ebf\u3002<\/p>\n<p>\u514d\u8d39\u4f53\u9a8c\uff1ahttps:\/\/chat.qwen.ai\/<\/p>\n<p>\u9b54\u642d\uff1ahttps:\/\/modelscope.cn\/collections\/Qwen3-Next-c314f23bd0264a<\/p>\n<p>HuggingFace\uff1ahttps:\/\/huggingface.co\/collections\/Qwen\/qwen3-next-68c25fd6838e585db8eeea9d<\/p>\n<p><a href=\"https:\/\/www.1ai.net\/en\/tag\/%e9%98%bf%e9%87%8c%e4%ba%91\" title=\"\u3010\u67e5\u770b\u542b\u6709[\u963f\u91cc\u4e91]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">\u963f\u91cc\u4e91<\/a>\u767e\u70bc\uff1ahttps:\/\/help.aliyun.com\/zh\/model-studio\/models#2c9c4628c9yyd<\/p>\n","protected":false},"excerpt":{"rendered":"<p>In the early hours of the morning, Alithong released the next generation of basic model structures, Qwen3-Next, and opened up the Qwen3-Next-80B-A3B series based on the architecture. According to official sources, it is believed that Context Length Scaling and Total Parameter Scaling are two major trends in the development of large models in the future. In order to further enhance the training and reasoning efficiency of the model under long context and large-scale general parameters, a new Qwen3-Next model structure was designed. It was described that the MoE model structure of Qwen3-Next, compared to Qwen3, had the following core improvements: M.M<\/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":[7588,219,334],"collection":[],"class_list":["post-42939","post","type-post","status-publish","format-standard","hentry","category-news","tag-7588","tag-219","tag-334"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/42939","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=42939"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/42939\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=42939"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=42939"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=42939"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=42939"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}