{"id":7053,"date":"2026-02-07T14:08:03","date_gmt":"2026-02-07T06:08:03","guid":{"rendered":"https:\/\/blog.deephour.com\/?p=7053"},"modified":"2026-02-07T14:08:04","modified_gmt":"2026-02-07T06:08:04","slug":"%f0%9f%8c%9f-ai%e6%98%a5%e6%99%9a%e5%bc%80%e6%bc%94%ef%bc%81claude-opus-4-6-vs-gpt-5-3-codex%ef%bc%8c%e8%bf%99%e6%b3%a2%e6%93%8d%e4%bd%9c%e7%9b%b4%e6%8e%a5%e8%ae%a9%e7%a8%8b%e5%ba%8f%e5%91%98%e9%9b%86","status":"publish","type":"post","link":"https:\/\/blog.deephour.com\/index.php\/2026\/02\/07\/%f0%9f%8c%9f-ai%e6%98%a5%e6%99%9a%e5%bc%80%e6%bc%94%ef%bc%81claude-opus-4-6-vs-gpt-5-3-codex%ef%bc%8c%e8%bf%99%e6%b3%a2%e6%93%8d%e4%bd%9c%e7%9b%b4%e6%8e%a5%e8%ae%a9%e7%a8%8b%e5%ba%8f%e5%91%98%e9%9b%86\/","title":{"rendered":"\ud83c\udf1f AI\u6625\u665a\u5f00\u6f14\uff01Claude Opus 4.6 vs GPT-5.3 Codex\uff0c\u8fd9\u6ce2\u64cd\u4f5c\u76f4\u63a5\u8ba9\u7a0b\u5e8f\u5458\u96c6\u4f53\u5931\u7720"},"content":{"rendered":"\n

<\/p>\n\n\n\n

\u5c31\u5728\u521a\u521a\uff0c\u7845\u8c37AI\u5708\u4e0a\u6f14\u4e86\u4e00\u573a”\u706b\u661f\u649e\u5730\u7403”\u7684\u79d1\u6280\u6625\u665a\uff01Anthropic\u548cOpenAI\u51e0\u4e4e\u540c\u65f6\u53d1\u5e03\u4e86\u81ea\u5df1\u7684\u738b\u724cAI\u6a21\u578b\u2014\u2014Claude Opus 4.6 \u548c GPT-5.3 Codex\u3002\u8fd9\u54ea\u662f\u66f4\u65b0\uff0c\u5206\u660e\u662f\u4e24\u5bb6\u5de8\u5934\u7684”\u4e2d\u95e8\u5bf9\u72d9”\uff0c\u706b\u836f\u5473\u76f4\u51b2\u5929\u7075\u76d6\uff01<\/p>\n\n\n\n

\ud83d\udd25 \u4e00\u573a\u7f16\u7a0b\u754c\u7684”\u5965\u65af\u5361”\u4e89\u593a\u6218<\/h2>\n\n\n\n

\u8bf4\u771f\u8bdd\uff0c\u8fd9\u4e2a\u6d88\u606f\u8ba9\u6211\u51cc\u6668\u4e09\u70b9\u624d\u7761\u7740\u3002\u6bd5\u7adf\uff0c\u6211\u4eec\u8fd9\u4e9b\u9760\u4ee3\u7801\u5403\u996d\u7684\u7a0b\u5e8f\u5458\uff0c\u5df2\u7ecf\u4e60\u60ef\u6027\u5730\u5c06GPT\u548cClaude\u5f53\u4f5c\u5de5\u5177\u7bb1\u91cc\u7684”\u4e07\u80fd\u87ba\u4e1d\u5200”\u3002\u4f46\u8fd9\u6b21\uff0c\u4e24\u5bb6\u5de8\u5934\u4e0d\u662f\u7b80\u5355\u5730\u5347\u7ea7\u4e86\u6a21\u578b\uff0c\u800c\u662f\u5728\u7f16\u7a0b\u80fd\u529b\u3001\u957f\u6587\u672c\u5904\u7406\u548cAI\u81ea\u4e3b\u8fdb\u5316<\/strong>\u8fd9\u4e09\u4e2a\u7ef4\u5ea6\u4e0a\u5f00\u6253\u4e86\u4e00\u573a\u786c\u4ed7\u3002<\/p>\n\n\n\n

Claude Opus 4.6\u5728Terminal-Bench 2.0\u6d4b\u8bd5\u4e2d\u5f97\u520665.4%\uff0c\u800cGPT-5.3 Codex\u5219\u4ee577.3%\u7684\u5f97\u5206\u5f3a\u52bf\u78be\u538b\u3002\u8fd9\u53ef\u4e0d\u662f\u5c0f\u5dee\u8ddd\uff0c\u800c\u662f\u8d28\u7684\u98de\u8dc3<\/strong>\uff01\u5728OSWorld-Verified\u6d4b\u8bd5\u4e2d\uff0cGPT-5.3 Codex\u5f97\u520664.7%\u3002<\/p>\n\n\n\n

\u66f4\u4ee4\u4eba\u60ca\u53f9\u7684\u662f\uff0cClaude Opus 4.6\u5728GDPval-AA<\/strong>\u8fd9\u4e2a\u7ecf\u6d4e\u4ef7\u503c\u5de5\u4f5c\u6d4b\u8bd5\u4e2d\uff0c\u4ee51606 Elo\u5206\u5927\u5e45\u9886\u5148GPT-5.2\uff081462\u5206\uff09\uff0c\u8fd9\u4e2a\u5dee\u8ddd\u76f8\u5f53\u4e8e\u670970%\u7684\u6982\u7387\u6218\u80dc\u5bf9\u624b\u3002\u800c\u5728ARC AGI 2<\/strong>\u8fd9\u79cd\u6d4b\u8bd5\u6d41\u4f53\u667a\u529b\u7684\u9879\u76ee\u4e0a\uff0cClaude Opus 4.6\u76f4\u63a5\u98d9\u523068.8%\uff0c\u51e0\u4e4e\u6478\u52307\u5b57\u5934\u4e86\u3002<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\ud83e\udde0 \u6df1\u5ea6\u601d\u8003 vs \u5feb\u901f\u6267\u884c<\/h2>\n\n\n\n

\u8fd9\u662f\u4e00\u573a\u601d\u7ef4\u6a21\u5f0f<\/strong>\u7684\u5bf9\u51b3\uff1aClaude Opus 4.6\u662f”\u9ad8\u4e0a\u9650\uff0c\u9ad8\u65b9\u5dee”\u7684\u5929\u624d\u578b\u9009\u624b\uff0c\u800cGPT-5.3 Codex\u5219\u662f”\u9ad8\u53ef\u9760\uff0c\u4f4e\u65b9\u5dee”\u7684\u5de5\u7a0b\u5e08\u578b\u9009\u624b\u3002<\/p>\n\n\n\n

Dan Shipper\u8fd9\u4f4d\u77e5\u540dAI\u8bc4\u6d4b\u4eba\u66fe\u8fd9\u6837\u5f62\u5bb9\uff1aClaude Opus 4.6\u80fd\u89e3\u51b3\u8ba9iOS\u56e2\u961f\u5361\u4e86\u4e24\u4e2a\u6708\u7684\u529f\u80fd\u96be\u9898\uff0c\u4f46\u5076\u5c14\u4f1a”\u8fc7\u5ea6\u81ea\u4fe1”\uff0c\u80e1\u8bf4\u516b\u9053\uff1b\u800cGPT-5.3 Codex\u867d\u7136\u5728\u521b\u9020\u6027\u4efb\u52a1\u4e0a\u7a0d\u900a\u4e00\u7b79\uff0c\u4f46”\u63a8\u7406\u901f\u5ea6\u63d0\u534725%\uff0c\u51e0\u4e4e\u4e0d\u72af\u4f4e\u7ea7\u9519\u8bef”\u3002<\/p>\n\n\n\n

\u6700\u8ba9\u6211\u773c\u524d\u4e00\u4eae\u7684\u662fGPT-5.3 Codex\u7684”\u81ea\u6211\u8fdb\u5316”\u80fd\u529b\u2014\u2014\u8fd9\u662fOpenAI\u9996\u6b21\u8ba9AI\u53c2\u4e0e\u81ea\u8eab\u5f00\u53d1\u3002\u5728\u771f\u5b9e\u6d4b\u8bd5\u4e2d\uff0cGPT-5.3 Codex\u4e0d\u4ec5\u80fd\u81ea\u5df1\u4fee\u590dbug\uff0c\u8fd8\u80fd\u5728\u51e0\u5c0f\u65f6\u5185\u6784\u5efa\u51fa\u5305\u542b\u591a\u5f20\u5730\u56fe\u7684\u8d5b\u8f66\u6e38\u620f\u548c\u6df1\u6d77\u6f5c\u6c34\u6e38\u620f\uff01<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\ud83d\udcbc \u8c01\u66f4\u9002\u5408\u4f60\uff1f\u6211\u7684\u5b9e\u6d4b\u5efa\u8bae<\/h2>\n\n\n\n

\u5982\u679c\u4f60\u662f\u9700\u8981\u5904\u7406\u5927\u578b\u4ee3\u7801\u5e93\u3001\u957f\u6587\u672c\u7406\u89e3\u7684\u5f00\u53d1\u8005<\/strong>\uff0cClaude Opus 4.6\u7edd\u5bf9\u662f\u9996\u9009\u3002\u5b83\u652f\u6301100\u4e07token\u4e0a\u4e0b\u6587\u7a97\u53e3\uff0c\u80fd\u5904\u7406\u76f8\u5f53\u4e8e75\u4e07\u82f1\u6587\u5355\u8bcd\u7684\u5185\u5bb9\u3002\u5728MRCR v2\u6d4b\u8bd5\u4e2d\uff0c\u5b83\u5728”\u5927\u6d77\u635e\u9488”\u4efb\u52a1\u4e0a\u5f97\u520676%\uff0c\u78be\u538b\u4e86Sonnet 4.5\u768418.5%\u3002<\/p>\n\n\n\n

\u800c\u5982\u679c\u4f60\u9700\u8981\u5feb\u901f\u8fed\u4ee3\u3001\u7a33\u5b9a\u4ea4\u4ed8<\/strong>\uff0cGPT-5.3 Codex\u4f1a\u662f\u66f4\u597d\u7684\u9009\u62e9\u3002\u5b83\u7528\u66f4\u5c11\u7684token\u5b8c\u6210\u4efb\u52a1\uff0c\u901f\u5ea6\u63d0\u534725%\uff0c\u5b8c\u7f8e\u9002\u5408\u65e5\u5e38\u7f16\u7801\u548c\u8fd0\u7ef4\u4efb\u52a1\u3002<\/p>\n\n\n\n

\u6700\u8ba9\u4eba\u9707\u64bc\u7684\u662f\uff0cGPT-5.3 Codex\u5728\u529e\u516c\u8f6f\u4ef6\u96c6\u6210<\/strong>\u65b9\u9762\u4e5f\u6709\u91cd\u5927\u7a81\u7834\u3002\u5b83\u80fd\u50cf\u771f\u4eba\u4e00\u6837\u64cd\u4f5cExcel\u548cPowerPoint\uff0c\u5728\u6784\u5efa\u7f51\u9875\u65f6\u81ea\u52a8\u5c06\u5e74\u5ea6\u8ba1\u5212\u6362\u7b97\u6210\u6253\u6298\u540e\u7684\u6708\u4ed8\u4ef7\u683c\uff0c\u751a\u81f3\u8d34\u5fc3\u5730\u81ea\u52a8\u8865\u5145\u7528\u6237\u8bc4\u4ef7\u8f6e\u64ad\u3002<\/p>\n\n\n\n

\ud83c\udf0d \u8c01\u5728\u80cc\u540e\u63a8\u52a8\u8fd9\u573aAI\u6625\u665a\uff1f<\/h2>\n\n\n\n

\u8fd9\u6b21\u53d1\u5e03\u4f1a\u7684\u80cc\u540e\uff0c\u6709\u7740\u66f4\u5b8f\u5927\u7684\u6218\u7565\u610f\u56fe\u3002\u4e24\u5bb6\u516c\u53f8\u90fd\u8ba1\u5212\u57282026\u5e74\u4e0b\u534a\u5e74\u52302027\u5e74<\/strong>\u542f\u52a8\u4e0a\u5e02\u7a0b\u5e8f\uff0c\u73b0\u5728\u6b63\u662f\u8bc1\u660e\u81ea\u5df1\u6280\u672f\u5b9e\u529b\u7684\u5173\u952e\u65f6\u523b\u3002<\/p>\n\n\n\n

\u5c31\u5728\u4e24\u5468\u524d\uff0c\u82f1\u4f1f\u8fbe\u521a\u5411Anthropic\u6295\u8d44100\u4ebf\u7f8e\u5143\uff0c\u6d88\u606f\u4f20\u51fa\u540e\u4e0d\u523072\u5c0f\u65f6\uff0c\u5c31\u5411OpenAI\u6ce8\u8d44200\u4ebf\u7f8e\u5143\u3002\u9ec4\u4ec1\u52cb\u7684\u7b97\u76d8\u6253\u5f97\u5f88\u6e05\u695a\uff1a\u4e24\u8fb9\u90fd\u62bc\u6ce8\uff0c\u8c01\u8d62\u90fd\u4e0d\u4e8f\u3002<\/p>\n\n\n\n

\u8fd9\u662f\u4e00\u573a\u6280\u672f\u8d44\u672c\u5316<\/strong>\u7684\u76db\u5bb4\u2014\u2014\u8c01\u7684\u6a21\u578b\u66f4\u5f3a\uff0c\u8c01\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\u66f4\u6709\u8bf4\u670d\u529b\uff0c\u8c01\u5c31\u80fd\u5728IPO\u65f6\u8981\u5230\u66f4\u9ad8\u7684\u4ef7\u683c\uff0c\u62ff\u5230\u66f4\u591a\u7684\u7b79\u7801\u3002<\/p>\n\n\n\n

\ud83c\udf1f \u4f60\u7684\u4f4d\u7f6e\u5728\u54ea\uff1f<\/h2>\n\n\n\n

\u6211\u4eec\u5df2\u7ecf\u8fdb\u5165\u4e86\u4e00\u4e2a\u65b0\u65f6\u4ee3\uff1aAI\u4e0d\u518d\u53ea\u662f\u8f85\u52a9\u5de5\u5177\uff0c\u800c\u662f\u80fd\u81ea\u4e3b\u601d\u8003\u3001\u534f\u4f5c\u3001\u8fdb\u5316<\/strong>\u7684\u540c\u4e8b\u7ea7\u4f19\u4f34\u3002\u5982\u679c\u4f60\u8fd8\u5728\u7528\u4e0a\u4e00\u4ee3AI\u6a21\u578b\uff0c\u53ef\u80fd\u5df2\u7ecf\u843d\u540e\u4e86\u3002<\/p>\n\n\n\n

\u4f46\u522b\u62c5\u5fc3\uff0c\u8fd9\u662f\u4e00\u573a\u5171\u8d62\u7684\u8f83\u91cf<\/strong>\u3002\u5c31\u50cf\u6211\u5e38\u8bf4\u7684\uff1a“\u6280\u672f\u5728\u8fdb\u6b65\uff0c\u4f46\u4eba\u7c7b\u7684\u667a\u6167\u6c38\u8fdc\u7ad9\u5728\u9876\u7aef\u3002”<\/strong><\/p>\n\n\n\n

\u4e0b\u6b21\u5f53\u4f60\u5728\u4ee3\u7801\u4e2d\u9047\u5230\u74f6\u9888\uff0c\u4e0d\u59a8\u8bd5\u8bd5\u8fd9\u4e24\u6b3e\u6a21\u578b\uff0c\u770b\u770b\u54ea\u4e2a\u80fd\u6210\u4e3a\u4f60\u7684”AI\u642d\u6863”\u3002<\/p>\n\n\n\n

AI\u4e0d\u4f1a\u53d6\u4ee3\u4eba\u7c7b\uff0c\u4f46\u4f1a\u6539\u53d8\u6211\u4eec\u4e0e\u5de5\u4f5c\u7684\u5173\u7cfb\u3002<\/strong><\/p>\n\n\n\n

\u4f60\u66f4\u5e0c\u671b\u62e5\u6709\u4e00\u4e2a\u80fd\u81ea\u7531\u53d1\u6325\u7684”\u5929\u624d”\uff0c\u8fd8\u662f\u4e00\u4e2a\u53ef\u9760\u7a33\u5b9a\u7684”\u8001\u9ec4\u725b”\uff1f\u6b22\u8fce\u8bc4\u8bba\u533a\u5206\u4eab\u4f60\u7684\u60f3\u6cd5\uff01\ud83d\udc4d<\/strong><\/p>\n","protected":false},"excerpt":{"rendered":"

\u5c31\u5728\u521a\u521a\uff0c\u7845\u8c37AI\u5708\u4e0a\u6f14\u4e86\u4e00\u573a”\u706b\u661f\u649e\u5730\u7403”\u7684\u79d1\u6280\u6625\u665a\uff01Anthropic\u548cOpenA… \u9605\u8bfb\u66f4\u591a »\ud83c\udf1f AI\u6625\u665a\u5f00\u6f14\uff01Claude Opus 4.6 vs GPT-5.3 Codex\uff0c\u8fd9\u6ce2\u64cd\u4f5c\u76f4\u63a5\u8ba9\u7a0b\u5e8f\u5458\u96c6\u4f53\u5931\u7720<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"neve_meta_sidebar":"","neve_meta_container":"","neve_meta_enable_content_width":"","neve_meta_content_width":0,"neve_meta_title_alignment":"","neve_meta_author_avatar":"","neve_post_elements_order":"","neve_meta_disable_header":"","neve_meta_disable_footer":"","neve_meta_disable_title":"","footnotes":""},"categories":[53,3],"tags":[10,11,28],"class_list":["post-7053","post","type-post","status-publish","format-standard","hentry","category-ai_tech-2","category-ai_tech","tag-ai","tag-claude","tag-openai"],"_links":{"self":[{"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/posts\/7053","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/comments?post=7053"}],"version-history":[{"count":1,"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/posts\/7053\/revisions"}],"predecessor-version":[{"id":7056,"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/posts\/7053\/revisions\/7056"}],"wp:attachment":[{"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/media?parent=7053"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/categories?post=7053"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.deephour.com\/index.php\/wp-json\/wp\/v2\/tags?post=7053"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}