{"id":55880,"date":"2026-08-16T17:14:29","date_gmt":"2026-08-16T09:14:29","guid":{"rendered":"https:\/\/www.1ai.net\/?p=55880"},"modified":"2026-08-16T17:14:29","modified_gmt":"2026-08-16T09:14:29","slug":"%e5%8d%83%e9%97%ae%e5%8a%9e%e5%85%ac%e9%a6%96%e5%8f%91%e4%b8%8a%e7%ba%bf-glm-5-3-%e5%92%8c-deepseek-v4-pro-%e6%a8%a1%e5%9e%8b%ef%bc%8c%e5%8f%af%e7%9b%b4%e6%8e%a5%e9%80%89%e7%94%a8","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/55880.html","title":{"rendered":"Start-up GLM-5.3 and DeepSeek V4 Pro models, directly available"},"content":{"rendered":"<p>On August 16th, the news was that recently<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%8d%83%e9%97%ae%e5%8a%9e%e5%85%ac\" title=\"_Other Organiser\" target=\"_blank\" >Office<\/a>Start on line\u00a0<strong>GM-5.3<\/strong>\u00a0and\u00a0<strong>DeepSeek V4Pro<\/strong>\u00a0Two front-line models. As of this date, the user can select the front-line model at the front page of the product\u3002<\/p>\n<p>DeepSeek V4 Pro is the latest DeepSeek flagship model<strong>support<\/strong>\u00a0<strong>1 million token context<\/strong>, the highest 384K output. The Agent task-linking capability was enhanced and ranked high on multiple long-range missions and automated mission benchmarks. Specialized tasks such as automation of workflows, suitable for complex programmes\u3002<\/p>\n<p>GLM-5.3 is the latest open source flagship model of the spectra. Based on the non-change base model, 50% performance was enhanced by further post-training, Scaling, compared to GLM-5.2, to better accommodate long-range complex engineering tasks; and the strong potential for cyber-security defensive scenes was demonstrated in security missions such as white box code review and gap detection\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-55881\" title=\"6bacb278jjuyt000sd000mn00em\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/08\/6bacb278j00tjuuyt000sd000mn00emp.jpg\" alt=\"6bacb278jjuyt000sd000mn00em\" width=\"815\" height=\"526\" \/><\/p>\n<p>1AI noted that Agent's \u201cthousands of questions\u201d product under the Ali Baba flag opened a public survey early this month. The application has a free, personal standard, personal advanced, etc. subscription package with specific subscription prices as follows:<\/p>\n<p>Personal Edition:<\/p>\n<ul>\n<li>Free version: $ 0\/ month<\/li>\n<li>Personal standard version: original 98\/month, 8 discount 78\/month<\/li>\n<li>Personal advanced version: original price 198 \/ month, 8 discount 158 \/ month<\/li>\n<\/ul>\n<p>Enterprise Edition:<\/p>\n<ul>\n<li>Standard enterprise version: $198\/month\/seat<\/li>\n<li>Enterprise flagship: need to consult official<\/li>\n<\/ul>\n<p>It was explained that the new \u201cthousands of questions\u201d were fully integrated with the products of QoderWork, MuleRun and Goku, and that it was the first in the industry to support desktop-end, cloud-end and business-cooperative Agent, and was committed to helping individuals and businesses to increase office productivity\u3002<\/p>","protected":false},"excerpt":{"rendered":"<p>On August 16, according to the news, in recent days, thousands of questions were sent on line GLM-5.3 and DeepSeek V4 Pro, two front-line models. As of this date, the user can select the front-line model at the front page of the product. DeepSeek V4 Pro is the latest DeepSeek flagship large model that supports a million token context, with the highest 384K output. The Agent task-linking capability was enhanced and ranked high on multiple long-range missions and automated mission benchmarks. Specialized tasks such as automation of workflows, suitable for complex programmes. GLM-5.3 is the latest open source flagship model of the spectra. On the basis of the base model, through further post-training Scaling<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[8724],"collection":[],"class_list":["post-55880","post","type-post","status-publish","format-standard","hentry","category-news","tag-8724"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/55880","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=55880"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/55880\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=55880"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=55880"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=55880"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=55880"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}