{"id":49774,"date":"2026-02-06T11:58:26","date_gmt":"2026-02-06T03:58:26","guid":{"rendered":"https:\/\/www.1ai.net\/?p=49774"},"modified":"2026-02-06T11:58:26","modified_gmt":"2026-02-06T03:58:26","slug":"openai-%e4%b8%8e-anthropic-%e5%90%8c%e6%97%b6%e5%8f%91%e5%b8%83%e9%87%8d%e7%a3%85%e6%a8%a1%e5%9e%8b%e6%9b%b4%e6%96%b0","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/49774.html","title":{"rendered":"OpenAI and Anthropic simultaneously publish heavy-pound model updates"},"content":{"rendered":"<p>2 \u6708 6 \u65e5\u6d88\u606f\uff0c\u4eca\u5929\u51cc\u6668\uff0c<a href=\"https:\/\/www.1ai.net\/en\/tag\/openai\" title=\"\u3010\u67e5\u770b\u542b\u6709[OpenAI]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">OpenAI<\/a> \u4e0e <a href=\"https:\/\/www.1ai.net\/en\/tag\/anthropic\" title=\"\u3010\u67e5\u770b\u542b\u6709[Anthropic]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">Anthropic<\/a> \u5148\u540e\u53d1\u5e03\u91cd\u78c5<a href=\"https:\/\/www.1ai.net\/en\/tag\/%e6%a8%a1%e5%9e%8b\" title=\"\u3010\u67e5\u770b\u542b\u6709[\u6a21\u578b]\u6807\u7b7e\u7684\u6587\u7ae0\u3011\" target=\"_blank\" rel=\"noopener\">\u6a21\u578b<\/a>\u66f4\u65b0\uff0c\u5206\u522b\u63a8\u51fa GPT\u20115.3\u2011Codex \u4e0e Claude Opus 4.6\u3002<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-49775\" title=\"12a1ac12j00ta0r09012td000u000y5m\" src=\"https:\/\/www.1ai.net\/wp-content\/uploads\/2026\/02\/12a1ac12j00ta0r09012td000u000y5m.jpg\" alt=\"12a1ac12j00ta0r09012td000u000y5m\" width=\"1080\" height=\"1229\" \/><\/p>\n<p>OpenAI \u5ba3\u79f0\uff0cGPT\u20115.3\u2011Codex \u662f\u5176\u9996\u4e2a\u5728\u81ea\u8eab\u7814\u53d1\u8fc7\u7a0b\u4e2d\u53d1\u6325\u5173\u952e\u4f5c\u7528\u7684\u6a21\u578b\uff0c\u56e2\u961f\u4f7f\u7528\u65e9\u671f\u7248\u672c\u53c2\u4e0e\u8c03\u8bd5\u8bad\u7ec3\u3001\u90e8\u7f72\u7ba1\u7406\u4e0e\u8bc4\u4f30\u5206\u6790\u3002<\/p>\n<p>\u8be5\u6a21\u578b\u5728 SWE\u2011Bench Pro\u3001Terminal\u2011Bench 2.0 \u4e0e OSWorld\u2011Verified \u7b49\u591a\u9879\u8bc4\u6d4b\u4e2d\u5237\u65b0\u7eaa\u5f55\uff0cTerminal\u2011Bench 2.0 \u5f97\u5206\u8fbe\u5230 77.3%\uff0c\u8f83 GPT\u20115.2 \u63d0\u5347\u663e\u8457\uff1b\u5728 OSWorld\u2011Verified \u4e2d\u51c6\u786e\u7387\u63d0\u5347\u81f3 64.7%\uff0c\u903c\u8fd1\u4eba\u7c7b\u5e73\u5747\u6c34\u5e73\u3002<\/p>\n<p>OpenAI \u5f3a\u8c03\uff0cGPT\u20115.3\u2011Codex \u4e0d\u4ec5\u80fd\u5199\u4ee3\u7801\uff0c\u8fd8\u80fd\u6267\u884c\u8de8\u8f6f\u4ef6\u7684\u590d\u6742\u64cd\u4f5c\uff0c\u652f\u6301\u957f\u65f6\u4efb\u52a1\u3001\u5b9e\u65f6\u4ea4\u4e92\uff0c\u5e76\u80fd\u5728\u6784\u5efa\u7f51\u9875\u3001\u6e38\u620f\u7b49\u4efb\u52a1\u4e2d\u5c55\u73b0\u66f4\u5f3a\u7684\u81ea\u4e3b\u6027\u4e0e\u610f\u56fe\u7406\u89e3\u80fd\u529b\u3002<\/p>\n<p>\u51e0\u4e4e\u540c\u4e00\u65f6\u95f4\uff0cAnthropic \u53d1\u5e03\u4e86 Claude Opus 4.6\uff0c\u4e3b\u6253\u63a8\u7406\u80fd\u529b\u3001\u53ef\u9760\u6027\u4e0e\u957f\u4e0a\u4e0b\u6587\u5904\u7406\u80fd\u529b\u3002\u5176\u9996\u6b21\u5f15\u5165 1M Token \u4e0a\u4e0b\u6587\u7a97\u53e3\uff08Beta\uff09\uff0c\u5728 MRCR v2 \u957f\u6587\u672c\u68c0\u7d22\u6d4b\u8bd5\u4e2d\u53ec\u56de\u7387\u8fbe\u5230 76%\uff0c\u8fdc\u8d85\u524d\u4ee3\u6a21\u578b\u3002<\/p>\n<p>Opus 4.6 \u5728 GDPval\u2011AA\uff08\u9ad8\u7ecf\u6d4e\u4ef7\u503c\u77e5\u8bc6\u5de5\u4f5c\u8bc4\u6d4b\uff09\u4e2d\u6bd4 GPT\u20115.2 \u9ad8\u51fa\u7ea6 144 Elo\uff0c\u5728 Humanity\u2019s Last Exam \u4e0e BrowseComp \u7b49\u8bc4\u6d4b\u4e2d\u540c\u6837\u9886\u5148\u3002<\/p>\n<p>Anthropic \u540c\u65f6\u63a8\u51fa Agent Teams \u529f\u80fd\uff0c\u5141\u8bb8\u591a\u4e2a\u667a\u80fd\u4f53\u5e76\u884c\u534f\u4f5c\uff0c\u652f\u6301\u4efb\u52a1\u62c6\u89e3\u3001\u72ec\u7acb\u4e0a\u4e0b\u6587\u4e0e\u667a\u80fd\u4f53\u95f4\u901a\u4fe1\u3002\u5b98\u65b9\u5c55\u793a\u4e2d\uff0c16 \u4e2a Opus 4.6 \u667a\u80fd\u4f53\u5728\u4e24\u5468\u5185\u81ea\u4e3b\u5b8c\u6210\u4e86\u4e00\u4e2a 10 \u4e07\u884c\u4ee3\u7801\u7684 C \u8bed\u8a00\u7f16\u8bd1\u5668\uff0c\u5e76\u6210\u529f\u7f16\u8bd1 Linux 6.9 \u5185\u6838\u3002<\/p>\n<p>\u5728\u751f\u4ea7\u529b\u573a\u666f\u4e2d\uff0cAnthropic \u5c06 Claude \u6df1\u5ea6\u6574\u5408\u8fdb Excel \u4e0e PowerPoint\uff0c\u53ef\u6839\u636e\u8868\u683c\u81ea\u52a8\u751f\u6210\u6392\u7248\u4e00\u81f4\u7684\u6f14\u793a\u6587\u7a3f\uff0c\u5e76\u5728 Claude Cowork \u4e2d\u6267\u884c\u591a\u4efb\u52a1\u534f\u4f5c\u3002<\/p>\n<p>TechCrunch \u6307\u51fa\uff0cOpus 4.6 \u7684 Agent Teams \u8ba9\u666e\u901a\u5f00\u53d1\u8005\u4e5f\u80fd\u4f53\u9a8c\u300c\u6307\u6325 AI \u56e2\u961f\u300d\u7684\u5de5\u4f5c\u65b9\u5f0f\u3002<\/p>\n<p>OpenAI \u5f3a\u8c03 GPT\u20115.3\u2011Codex \u7684\u9ad8\u53ef\u9760\u6027\u3001\u4f4e\u65b9\u5dee\u7279\u6027\uff0c\u66f4\u9002\u5408\u5de5\u7a0b\u6267\u884c\u4e0e\u8fd0\u7ef4\uff1bAnthropic \u5219\u5f3a\u8c03 Opus 4.6 \u7684\u9ad8\u4e0a\u9650\u63a8\u7406\u80fd\u529b\u4e0e\u957f\u6587\u672c\u5904\u7406\u4f18\u52bf\uff0c\u66f4\u9002\u5408\u91d1\u878d\u3001\u6cd5\u5f8b\u4e0e\u590d\u6742\u51b3\u7b56\u573a\u666f\u3002<\/p>\n<p>\u5f6d\u535a\u793e\u62a5\u9053\u6307\u51fa\uff0cOpus 4.6 \u7684\u53d1\u5e03\u5df2\u5bf9\u91d1\u878d\u6570\u636e\u670d\u52a1\u884c\u4e1a\u9020\u6210\u51b2\u51fb\uff0c\u591a\u5bb6\u4e0a\u5e02\u516c\u53f8\u80a1\u4ef7\u77ed\u7ebf\u4e0b\u8dcc\u3002<\/p>\n<p>\u503c\u5f97\u6ce8\u610f\u7684\u662f\uff0c\u636e TechCrunch \u62a5\u9053\uff0cOpenAI \u7684 GPT\u20115.3\u2011Codex \u539f\u8ba1\u5212\u4e0e Anthropic \u540c\u65f6\u53d1\u5e03\uff0c\u4f46\u5728\u5bf9\u65b9\u63d0\u524d 15 \u5206\u949f\u516c\u5e03\u540e\uff0cOpenAI \u968f\u5373\u8c03\u6574\u53d1\u5e03\u65f6\u95f4\u5e76\u8fc5\u901f\u4e0a\u7ebf\u66f4\u65b0\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>In the morning of February 6th, OpenAI and Anthropic released an update of the heavy-pound model, launching GPT \u2013 5.3 \u2013 Codex and Claude Opus 4.6, respectively. OpenAI claims that GPT-5.3-Codex was its first model to play a key role in its own research and development, and that the team used an early version to participate in commissioning training, deployment management and evaluation analysis. The model was updated in a number of assessments, such as SWE-Bench Pro, Terminal-Bench 2.0 and OSWorld-Verified, with a score of 773% compared to GPT-5<\/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":[320,190,1489],"collection":[],"class_list":["post-49774","post","type-post","status-publish","format-standard","hentry","category-news","tag-anthropic","tag-openai","tag-1489"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/49774","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=49774"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/49774\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=49774"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=49774"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=49774"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=49774"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}