{"id":30743,"date":"2025-03-15T11:27:49","date_gmt":"2025-03-15T03:27:49","guid":{"rendered":"https:\/\/www.1ai.net\/?p=30743"},"modified":"2025-03-15T11:27:49","modified_gmt":"2025-03-15T03:27:49","slug":"%e5%8a%a0%e6%8b%bf%e5%a4%a7%e5%88%9d%e5%88%9b%e5%85%ac%e5%8f%b8%e6%8e%a8%e5%87%ba-command-a-%e8%bd%bb%e9%87%8f%e7%ba%a7-ai-%e6%a8%a1%e5%9e%8b%ef%bc%8c%e5%8f%b7%e7%a7%b0%e4%bb%85%e9%9c%80%e4%b8%a4","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/30743.html","title":{"rendered":"Canadian Startup Launches Command A Lightweight AI Model, Claims to Require Only Two NVIDIA A100 \/ H100 GPUs for Deployment"},"content":{"rendered":"<p><a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%8a%a0%e6%8b%bf%e5%a4%a7\" title=\"\u3010\u67e5\u770b\u542b\u6709[\u52a0\u62ff\u5927]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">\u52a0\u62ff\u5927<\/a> AI <a href=\"https:\/\/www.1ai.net\/en\/tag\/%e5%88%9d%e5%88%9b%e5%85%ac%e5%8f%b8\" title=\"\u3010\u67e5\u770b\u542b\u6709[\u521d\u521b\u516c\u53f8]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">\u521d\u521b\u516c\u53f8<\/a> Cohere \u53d1\u5e03\u4e86\u4e00\u6b3e\u540d\u4e3a\u201c<a href=\"https:\/\/www.1ai.net\/en\/tag\/command-a\" title=\"\u3010\u67e5\u770b\u542b\u6709[Command A]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">Command A<\/a>\u201d\u7684 <a href=\"https:\/\/www.1ai.net\/en\/tag\/ai%e6%a8%a1%e5%9e%8b\" title=\"\u3010\u67e5\u770b\u542b\u6709[AI\u6a21\u578b]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">AI\u6a21\u578b<\/a>\uff0c\u8be5\u6a21\u578b\u4e3b\u6253\u8f7b\u91cf\u5316\u5e94\u7528\uff0c\u53f7\u79f0\u53ea\u9700\u8981\u4e24\u5757\u82f1\u4f1f\u8fbe A100 \u6216 H100 GPU \u5373\u53ef\u8f7b\u677e\u90e8\u7f72\uff0c\u53f7\u79f0\u201c\u8868\u73b0\u80fd\u548c GPT-4o \u76f8\u5ab2\u7f8e\u201d\uff0c\u5b9e\u73b0\u201c\u7528\u6700\u5c11\u7684\u786c\u4ef6\u53d1\u6325\u6700\u5927\u7684\u6027\u80fd\u201d\u3002<\/p>\n<p><img decoding=\"async\" id=\"netease1742009186715\" contenteditable=\"false\" src=\"http:\/\/dingyue.ws.126.net\/2025\/0315\/823b5f0ej00st5ax0001ud000v900qep.jpg\" \/><\/p>\n<p>Cohere \u8868\u793a\uff0cCommand A \u4e13\u95e8\u4e3a\u4e2d\u5c0f\u578b\u4f01\u4e1a\u73af\u5883\u8bbe\u8ba1\uff0c<strong>\u5176\u652f\u6301 256k \u7684\u4e0a\u4e0b\u6587\u957f\u5ea6\u548c 23 \u79cd\u8bed\u8a00<\/strong>\uff0c\u4f5c\u4e3a\u6bd4\u8f83\uff0c\u5176\u4ed6\u7ade\u4e89\u5bf9\u624b\u7684\u201c\u540c\u7c7b\u6a21\u578b\u201d\u9700\u8981 32 \u5757 GPU \u624d\u80fd\u90e8\u7f72\u3002<\/p>\n<p>\u5728\u6027\u80fd\u6d4b\u8bd5\u4e2d\uff0c<strong>Command A \u6700\u5feb\u6bcf\u79d2\u53ef\u4ee5\u8f93\u51fa 156 \u4e2a Token<\/strong>\uff0c\u636e\u79f0\u6bd4 GPT-4o \u5feb 1.75 \u500d\u3002\u800c\u5728\u6307\u4ee4\u8ffd\u8e2a\u3001SQL\u3001\u4ee3\u7406\u7a0b\u5e8f\u548c\u5de5\u5177\u4efb\u52a1\u7684\u57fa\u51c6\u6d4b\u8bd5\u4e0a\uff0cCommand A \u4e5f\u8868\u73b0\u51fa\u8272\u3002<\/p>\n<p>Cohere \u63f4\u5f15\u76f8\u5e94\u6027\u80fd\u6570\u636e\uff0c\u58f0\u79f0\u4e1a\u754c\u7684\u5927\u8bed\u8a00\u6a21\u578b\u5982\u679c\u201c\u89c4\u6a21\u8fc7\u5927\u201d\uff0c\u53cd\u800c\u5728\u8f93\u51fa\u7ed3\u679c\u65f6\u53ef\u80fd\u4f1a\u51fa\u73b0\u4e25\u91cd\u7684\u5ef6\u8fdf\u95ee\u9898\uff1b\u5982\u679c\u53ea\u662f\u60f3\u5feb\u901f\u5f97\u5230\u6b63\u786e\u7b54\u6848\uff0cCommand A \u76f8\u5bf9\u662f\u4e2a\u4e0d\u9519\u7684\u9009\u62e9\u3002<\/p>\n<p>\u76ee\u524d\uff0cCohere \u5df2\u5c06\u76f8\u5e94 Command A \u53d1\u5e03\u5728\u00a0Hugging Face \u5e73\u53f0\uff08<a href=\"https:\/\/huggingface.co\/CohereForAI\/c4ai-command-a-03-2025?ref=cohere-ai.ghost.io\">\u70b9\u6b64\u8bbf\u95ee<\/a>\uff09\uff0c\u5bf9\u5b66\u672f\u754c\u5f00\u653e\u4f7f\u7528\uff0c\u672a\u6765\u8fd8\u4f1a\u767b\u9646\u5176\u4ed6\u4e91\u670d\u52a1\u5e73\u53f0\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Canadian AI startup Cohere has released an AI model called \"Command A\", which focuses on lightweight applications and can be easily deployed with just two NVIDIA A100 or H100 GPUs, claiming \"performance comparable to GPT-4o\". Cohere said Command is a lightweight application that can be easily deployed with just two NVIDIA A100 or H100 GPUs, claiming \"performance comparable to GPT-4o\" and \"maximum performance with minimum hardware. Designed specifically for SMB environments, Command A supports 256k context length and 23 languages, Cohere said, in comparison to competitors' \"comparable models\" that require 32 GPUs to deploy. In performance tests, Command A could output up to 156 tokens per second, according to the company.<\/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":[167,5982,309,2896],"collection":[],"class_list":["post-30743","post","type-post","status-publish","format-standard","hentry","category-news","tag-ai","tag-command-a","tag-309","tag-2896"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/30743","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=30743"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/30743\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=30743"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=30743"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=30743"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=30743"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}