{"id":58855,"date":"2026-10-07T13:51:00","date_gmt":"2026-10-07T05:51:00","guid":{"rendered":"https:\/\/www.1ai.net\/?p=58855"},"modified":"2026-10-07T13:51:00","modified_gmt":"2026-10-07T05:51:00","slug":"%e5%ae%a3%e7%a7%b0%e6%ac%a7%e7%be%8e%e6%9c%80%e5%bc%ba%e5%bc%80%e6%ba%90%e6%a8%a1%e5%9e%8b%ef%bc%9amistral-ai-%e5%8f%91%e5%b8%83-mistral-large-4-%e5%85%ac%e5%bc%80%e9%a2%84%e8%a7%88","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/58855.html","title":{"rendered":"Proclaimed \"the most powerful open-source model in Europe and America\": Mistral AI launched the Mistral Large 4 open preview version at the end of the month"},"content":{"rendered":"

October 7, local time, October 6, French company Mistral<\/a> AI officially launched the public preview of the Mistral Large 4 model, short ML4, code name le Chonk\u3002<\/p>\n

As of this date, the user will be able to experience the preview API on the Mistral Studio, and the model weight will be officially available for download at the end of this month\u3002<\/p>\n

Officially, the top network security agencies, trusted partners and government regulators are jointly conducting a Red-Teaming test of the model in a real business environment. The relevant partners will be given specific test access to the model; the version of the test appropriately eases the content security clearance strategy and enhances the technical capability of cybersecurity\u3002<\/p>\n

\"e9f08f25j00tmiw7v007id000gk009bp\"<\/p>\n

ML4 is an original multimodular model with a total of 1 trillion and 49 billion active parameters, which is actually measured not only in the best sense of the worldOpen Source Model<\/a>,More than any open weight model developed in Europe and America<\/strong>.<\/p>\n

\n
    \n
  • IN KEY ENTERPRISE-LEVEL WORKLOADS SUCH AS CYBERSECURITY, FINANCE AND LAW, THE MODEL REACHES SOTA IN ALL OPEN SOURCE MODELS\u3002<\/li>\n
  • In certain specific areas (e.g. visual positioning), its performance goes even further and goes beyond current front-line closed-source models\u3002<\/li>\n<\/ul>\n<\/blockquote>\n

    ML4 was described as being within the European-based Self-Construction Data Centre of Mistral, based on 3,800 British Weeds Grace Blackwell GPU, which had been trained from scratch, with its European deployment programme being operated from the end to the end, in strict compliance with EU legislation and not subject to any other third-party digital service provider\u3002<\/p>\n

    ML4 TRAINING DATA HAVE A SIGNIFICANT MULTILINGUAL CHARACTER, WITH TEXT COVERAGE OF OVER 160 NATURAL LANGUAGES AND A COMPREHENSIVE COVERAGE OF ALL OFFICIAL LANGUAGES OF THE EU\u3002<\/p>\n

    On the price side, ML4 enters US$ 1.36 per million tokens (current exchange rate is approximately NMB 9.1) and US$ 4.18 per million tokens (current exchange rate is approximately NMB 28.1)\u3002<\/p>","protected":false},"excerpt":{"rendered":"

    News from 7 October, local time, 6 October, the French company Mistral AI officially launched the public preview of the Mistral Large 4 model, short ML4, code name le Chonk. As of this date, the user will be able to experience the preview API on the Mistral Studio, and the model weight will be officially available for download at the end of this month. Officially, the top network security agencies, trusted partners and government regulators are jointly conducting a Red-Teaming test of the model in a real business environment. The relevant partners will be given specific test access to the model; the version of the test will be appropriately relaxed for content security reviews<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[146],"tags":[559,862],"collection":[],"class_list":["post-58855","post","type-post","status-publish","format-standard","hentry","category-news","tag-mistral","tag-862"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/58855","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=58855"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/58855\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=58855"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=58855"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=58855"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=58855"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}