{"id":5065,"date":"2025-03-06T18:00:57","date_gmt":"2025-03-06T09:00:57","guid":{"rendered":"https:\/\/github.blog\/jp\/?p=5065"},"modified":"2025-03-06T17:56:03","modified_gmt":"2025-03-06T08:56:03","slug":"ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot","status":"publish","type":"post","link":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/","title":{"rendered":"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5"},"content":{"rendered":"<div class=\"col-12 offset-lg-1 col-lg-10\">\n<div class=\"text-mono f5-mktg color-fg-muted mb-12px\">Author<\/div>\n<div class=\"d-flex flex-wrap pb-1 flex-items-start\">\n<div class=\"d-flex flex-wrap\">\n<div class=\"author-block__pic mr-3\" style=\"margin-left: 30px;\"><img loading=\"lazy\" decoding=\"async\" class=\"avatar avatar-35 photo avatar-default\" src=\"https:\/\/avatars.githubusercontent.com\/u\/1358951?v=4&amp;s=35\" srcset=\"https:\/\/avatars.githubusercontent.com\/u\/1358951?v=4&amp;s=70 2x\" alt=\"Image of Connor Adams\" width=\"35\" height=\"35\" \/><\/div>\n<p class=\"hero-post__author-name f5 m-0\"><a href=\"https:\/\/github.blog\/author\/connorbadams\/\">Connor Adams<\/a><\/p>\n<div class=\"author-block__pic mr-3\" style=\"margin-left: 30px;\"><img loading=\"lazy\" decoding=\"async\" class=\"avatar avatar-35 photo avatar-default\" src=\"https:\/\/avatars.githubusercontent.com\/u\/957053?v=4&amp;s=35\" srcset=\"https:\/\/avatars.githubusercontent.com\/u\/957053?v=4&amp;s=70 2x\" alt=\"Image of Connor Adams\" width=\"35\" height=\"35\" \/><\/div>\n<p class=\"hero-post__author-name f5 m-0\"><a href=\"https:\/\/github.blog\/author\/klintron\/\">Klint Finley<\/a><\/p>\n<\/div>\n<\/div>\n<\/div>\n<hr \/>\n<p>\u6700\u8fd1\u3067\u306f\u3001AI\u30e2\u30c7\u30eb\u306e\u9078\u629e\u80a2\u304c\u5897\u3048\u307e\u3057\u305f\u3002OpenAI\u3001Google\u3001Anthropic\u3068\u3044\u3063\u305f\u30d7\u30ed\u30d7\u30e9\u30a4\u30a8\u30bf\u30ea\u306a\u57fa\u76e4\u30e2\u30c7\u30eb\u304b\u3089\u3001Meta\u3084Mistral\u306e\u3088\u3046\u306a\u5c0f\u898f\u6a21\u3067\u30aa\u30fc\u30d7\u30f3\u306a\u30aa\u30d7\u30b7\u30e7\u30f3\u307e\u3067\u3002\u6700\u65b0\u306e\u30e2\u30c7\u30eb\u3092\u8a66\u3057\u305f\u304f\u306a\u308a\u307e\u3059\u306d\u3002\u3057\u304b\u3057\u3001\u30e2\u30c7\u30eb\u304c\u65b0\u3057\u3044\u304b\u3089\u3068\u3044\u3063\u3066\u3001\u3042\u306a\u305f\u306e\u30e6\u30fc\u30b9\u30b1\u30fc\u30b9\u306b\u3068\u3063\u3066\u3088\u308a\u826f\u3044\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u767a\u63ee\u3059\u308b\u3068\u306f\u9650\u308a\u307e\u305b\u3093\u3002<\/p>\n<p>GitHub\u3067\u306f\u3001Anthropic\u306eClaude 3.7 Sonnet\u3001Google\u306eGemini 2.0 Flash\u3001OpenAI\u306eo2-mini\u3084GPT-4.5\u306a\u3069\u306e\u30b5\u30dd\u30fc\u30c8\u3092\u8ffd\u52a0\u3057\u3001<a href=\"https:\/\/github.blog\/changelog\/2025-02-26-openai-o3-mini-and-google-gemini-2-0-flash-now-available-for-copilot-in-jetbrains-ides\/\">GitHub Copilot\u3067\u5229\u7528\u3067\u304d\u308b\u30e2\u30c7\u30eb\u3092\u62e1\u5927<\/a>\u3057\u307e\u3057\u305f\u3002GitHub Copilot\u306b\u30e2\u30c7\u30eb\u3092\u8ffd\u52a0\u3059\u308b\u969b\u3001\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3001\u54c1\u8cea\u3001\u5b89\u5168\u6027\u3092\u5e38\u306b\u6700\u512a\u5148\u3057\u3066\u3044\u307e\u3059\u3002\u3053\u306e\u8a18\u4e8b\u3067\u306f\u3001GitHub Copilot\u30c1\u30fc\u30e0\u304cAI\u30e2\u30c7\u30eb\u3092\u8a55\u4fa1\u3057\u305f\u7d4c\u9a13\u3092\u3001\u30aa\u30d5\u30e9\u30a4\u30f3\u8a55\u4fa1\uff08\u672c\u756a\u74b0\u5883\u306b\u5909\u66f4\u3092\u52a0\u3048\u308b<em>\u524d\u306b<\/em>\u884c\u3046\u30c6\u30b9\u30c8\uff09\u306b\u7126\u70b9\u3092\u5f53\u3066\u3066\u7d39\u4ecb\u3057\u307e\u3059\u3002\u79c1\u305f\u3061\u306e\u7d4c\u9a13\u304c\u7686\u3055\u3093\u306e\u8a55\u4fa1\u306e\u53c2\u8003\u306b\u306a\u308c\u3070\u5e78\u3044\u3067\u3059\u3002<\/p>\n<h2 id=\"what-is-ai-model-evaluation\">AI\u30e2\u30c7\u30eb\u306e\u8a55\u4fa1\u3068\u306f\uff1f<\/h2>\n<p>AI\u30e2\u30c7\u30eb\u3068\u306f\u3001\u30b3\u30fc\u30c9\u3001\u30a2\u30eb\u30b4\u30ea\u30ba\u30e0\u3001\u5b66\u7fd2\u30c7\u30fc\u30bf\u3092\u7d44\u307f\u5408\u308f\u305b\u3066\u3001\u4eba\u9593\u306e\u77e5\u80fd\u3092\u4f55\u3089\u304b\u306e\u65b9\u6cd5\u3067\u30b7\u30df\u30e5\u30ec\u30fc\u30c8\u3059\u308b\u30b7\u30b9\u30c6\u30e0\u306e\u3053\u3068\u3067\u3059\u3002GitHub Copilot\u3067\u306f\u3001\u5927\u898f\u6a21\u8a00\u8a9e\u30e2\u30c7\u30eb\uff08Large Language Model\u3001LLM\uff09\u3068\u3057\u3066\u77e5\u3089\u308c\u308b\u3001\u4eba\u9593\u306e\u8a00\u8a9e\u306b\u57fa\u3065\u3044\u3066\u8a13\u7df4\u3055\u308c\u305f\u591a\u6570\u306eAI\u30e2\u30c7\u30eb\u3092\u4f7f\u7528\u3059\u308b\u9078\u629e\u80a2\u3092\u30e6\u30fc\u30b6\u30fc\u306b\u63d0\u4f9b\u3057\u3066\u3044\u307e\u3059\u3002OpenAI\u306e\u30e2\u30c7\u30eb\u306f\u3001ChatGPT\u306e\u4eba\u6c17\u306b\u3088\u308a\u6700\u3082\u3088\u304f\u77e5\u3089\u308c\u305fLLM\u306e\u4e00\u90e8\u3067\u3059\u304c\u3001Claude\u3001Gemini\u3001Meta\u306eLlama\u30e2\u30c7\u30eb\u306a\u3069\u306e\u4ed6\u306e\u30e2\u30c7\u30eb\u3082\u4eba\u6c17\u304c\u9ad8\u307e\u3063\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>AI\u30e2\u30c7\u30eb\u306e\u8a55\u4fa1\u3068\u306f\u3001\u3053\u308c\u3089\u306e\u30e2\u30c7\u30eb\u306e\u6027\u80fd\u3001\u54c1\u8cea\u3001\u5b89\u5168\u6027\u3092\u8a55\u4fa1\u3059\u308b\u3053\u3068\u3067\u3059\u3002\u8a55\u4fa1\u306b\u306f\u3001\u81ea\u52d5\u30c6\u30b9\u30c8\u3068\u624b\u52d5\u4f7f\u7528\u306e\u4e21\u65b9\u304c\u3042\u308a\u3001\u5229\u7528\u53ef\u80fd\u306a\u30d5\u30ec\u30fc\u30e0\u30ef\u30fc\u30af\u3084\u30d9\u30f3\u30c1\u30de\u30fc\u30af\u30c4\u30fc\u30eb\u306f\u6570\u591a\u304f\u3042\u308a\u307e\u3059\u304c\u3001\u79c1\u305f\u3061\u306f\u72ec\u81ea\u306e\u8a55\u4fa1\u30c4\u30fc\u30eb\u3092\u69cb\u7bc9\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u81ea\u52d5\u30c6\u30b9\u30c8\u3092\u6d3b\u7528\u3059\u308b\u3053\u3068\u3067\u3001\u5927\u898f\u6a21\u306a\u8a55\u4fa1\u304c\u53ef\u80fd\u306b\u306a\u308a\u3001\u591a\u6570\u306e\u30bf\u30b9\u30af\u306b\u5bfe\u3057\u3066\u30e2\u30c7\u30eb\u304c\u3069\u306e\u7a0b\u5ea6\u3046\u307e\u304f\u6a5f\u80fd\u3059\u308b\u304b\u3092\u898b\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002\u6b20\u70b9\u306f\u3001\u51fa\u529b\u3092\u8a55\u4fa1\u3059\u308b\u305f\u3081\u306b\u5ba2\u89b3\u7684\u306a\u57fa\u6e96\u304c\u5fc5\u8981\u306a\u5834\u5408\u304c\u3042\u308b\u3053\u3068\u3067\u3059\u3002\u624b\u52d5\u30c6\u30b9\u30c8\u306f\u3001\u30a2\u30a6\u30c8\u30d7\u30c3\u30c8\u306e\u54c1\u8cea\u3068\u7cbe\u5ea6\u3092\u3088\u308a\u4e3b\u89b3\u7684\u306b\u8a55\u4fa1\u3067\u304d\u307e\u3059\u304c\u3001\u6642\u9593\u304c\u304b\u304b\u308a\u307e\u3059\u3002\u3053\u306e\u4e8c\u3064\u306e\u30a2\u30d7\u30ed\u30fc\u30c1\u3092\u7d44\u307f\u5408\u308f\u305b\u308b\u3053\u3068\u3067\u3001\u81a8\u5927\u306a\u6642\u9593\u3092\u304b\u3051\u3066\u7121\u6570\u306e\u56de\u7b54\u3092\u624b\u4f5c\u696d\u3067\u7cbe\u67fb\u3057\u305f\u308a\u3001\u500b\u3005\u306e\u4e8b\u4f8b\u306b\u983c\u308a\u3059\u304e\u305f\u308a\u3059\u308b\u3053\u3068\u306a\u304f\u3001\u56de\u7b54\u306e\u4e3b\u89b3\u7684\u306a\u8cea\u3092\u628a\u63e1\u3067\u304d\u308b\u3088\u3046\u306b\u306a\u308a\u307e\u3059\u3002<\/p>\n<aside class=\"p-4 p-md-6 post-aside--large\" style=\"background-color:#f6f8fa; margin:30px 0px;\">\n<p style=\"font-size:20px; font-weight:bold;\">\u8cac\u4efb\u3042\u308bAI\u958b\u767a<\/p>\n<p>\u4f7f\u7528\u3059\u308b\u30e2\u30c7\u30eb\u306b\u304b\u304b\u308f\u3089\u305a\u3001GitHub Copilot\u306f\u30d7\u30ed\u30f3\u30d7\u30c8\u3068\u56de\u7b54\u306e\u4e21\u65b9\u3067\u3001\u30b3\u30fc\u30c9\u306b\u95a2\u9023\u3057\u306a\u3044\u8cea\u554f\u3001\u30d8\u30a4\u30c8\u30b9\u30d4\u30fc\u30c1\u3001\u6027\u7684\u306a\u5185\u5bb9\u3001\u66b4\u529b\u3001\u81ea\u50b7\u884c\u70ba\u306e\u8a3c\u62e0\u306a\u3069\u306e\u6709\u5bb3\u306a\u8a00\u8449\u306a\u3069\u306e\u95a2\u9023\u6027\u3092\u30c6\u30b9\u30c8\u3057\u307e\u3059\u3002\u79c1\u305f\u3061\u306f\u3001GitHub Copilot\u304c\u30e2\u30c7\u30eb\u3078\u306e\u8ab9\u8b17\u4e2d\u50b7\u3084\u4e0b\u54c1\u306a\u8a00\u8449\u3001\u30b3\u30fc\u30c9\u3084\u30bd\u30d5\u30c8\u30a6\u30a7\u30a2\u958b\u767a\u306b\u95a2\u4fc2\u306e\u306a\u3044\u8cea\u554f\u3068\u95a2\u308f\u308b\u3053\u3068\u3092\u671b\u3093\u3067\u3044\u307e\u305b\u3093\u3002\u540c\u6642\u306b\u3001\u79c1\u305f\u3061\u306f\u8fc5\u901f\u306a\u30cf\u30c3\u30ad\u30f3\u30b0\u304b\u3089\u3082\u5b88\u3063\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>GPT-4o\u304b\u3089Claude 3.5 Sonnet\u3001Gemini 2.0 Flash\u30d7\u30ed\u307e\u3067\u3001\u79c1\u305f\u3061\u304c\u5b9f\u88c5\u3059\u308b\u3059\u3079\u3066\u306e\u30e2\u30c7\u30eb\u306f\u3001\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u8a55\u4fa1\u3057\u3001\u79c1\u305f\u3061\u306e\u57fa\u6e96\u3092\u6e80\u305f\u3057\u3066\u3044\u308b\u304b\u3069\u3046\u304b\u3092\u5fb9\u5e95\u7684\u306b\u541f\u5473\u3055\u308c\u307e\u3059\u3002\u79c1\u305f\u3061\u306e\u8cac\u4efb\u3042\u308bAI\u8a55\u4fa1\u3068\u30ec\u30c3\u30c9\u30c1\u30fc\u30e0\u30c6\u30b9\u30c8\u306f\u5e83\u7bc4\u56f2\u306b\u53ca\u3073\u3001\u79c1\u305f\u3061\u304c\u54c1\u8cea\u3068\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u8a55\u4fa1\u3059\u308b\u305f\u3081\u306b\u4f7f\u7528\u3057\u3066\u3044\u308b\u306e\u3068\u540c\u3058\u30c6\u30af\u30cb\u30c3\u30af\u306e\u591a\u304f\u3092\u5b89\u5168\u6027\u306e\u8a55\u4fa1\u306b\u3082\u4f7f\u7528\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u8a73\u3057\u304f\u306f<a href=\"https:\/\/github.com\/trust-center\">GitHub Trust Center\u3092<\/a>\u3054\u89a7\u304f\u3060\u3055\u3044\u3002<\/p>\n<\/aside>\n<h2 id=\"automating-code-quality-tests\">\u30b3\u30fc\u30c9\u54c1\u8cea\u30c6\u30b9\u30c8\u306e\u81ea\u52d5\u5316<\/h2>\n<p>\u79c1\u305f\u3061\u306f4,000\u4ee5\u4e0a\u306e\u30aa\u30d5\u30e9\u30a4\u30f3\u30c6\u30b9\u30c8\u3092\u5b9f\u884c\u3057\u3066\u304a\u308a\u3001\u305d\u306e\u307b\u3068\u3093\u3069\u306f\u81ea\u52d5\u5316\u3055\u308c\u305fCI\u30d1\u30a4\u30d7\u30e9\u30a4\u30f3\u306e\u4e00\u90e8\u3068\u3057\u3066\u5b9f\u884c\u3055\u308c\u3066\u3044\u307e\u3059\u3002\u307e\u305f\u3001\u30ab\u30ca\u30ea\u30a2\u30c6\u30b9\u30c8\u306b\u4f3c\u305f\u5185\u90e8\u8a55\u4fa1\u3082\u30e9\u30a4\u30d6\u3067\u5b9f\u65bd\u3057\u3066\u304a\u308a\u3001\u591a\u6570\u306e Hubber(GitHub\u793e\u54e1) \u304c\u65b0\u3057\u3044\u30e2\u30c7\u30eb\u3092\u4f7f\u7528\u3067\u304d\u308b\u3088\u3046\u306b\u3057\u3066\u3044\u307e\u3059\u3002\u65b0\u3057\u3044\u30e2\u30c7\u30eb\u306e\u53ef\u80fd\u6027\u3060\u3051\u3067\u306a\u304f\u3001GitHub Copilot\u306e\u4e3b\u8981\u306a\u5909\u66f4\u306f\u3059\u3079\u3066\u3053\u306e\u65b9\u6cd5\u3067\u30c6\u30b9\u30c8\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<p>\u3053\u3053\u3067\u306f\u30aa\u30d5\u30e9\u30a4\u30f3\u30c6\u30b9\u30c8\u306b\u7126\u70b9\u3092\u5f53\u3066\u307e\u3059\u3002\u4f8b\u3048\u3070\u3001\u30b3\u30fc\u30c9\u30d9\u30fc\u30b9\u3092\u8a55\u4fa1\u30fb\u4fee\u6b63\u3059\u308b\u80fd\u529b\u306b\u3088\u3063\u3066\u3001\u5019\u88dc\u3068\u306a\u308b\u30e2\u30c7\u30eb\u3092\u8a55\u4fa1\u3057\u307e\u3059\u3002\u79c1\u305f\u3061\u306f\u3001CI\u30c6\u30b9\u30c8\u306b\u5408\u683c\u3057\u305f\u7d04100\u306e\u30b3\u30f3\u30c6\u30ca\u5316\u3055\u308c\u305f\u30ea\u30dd\u30b8\u30c8\u30ea\u306e\u30b3\u30ec\u30af\u30b7\u30e7\u30f3\u3092\u6301\u3063\u3066\u3044\u307e\u3059\u3002\u3053\u308c\u3089\u306e\u30ea\u30dd\u30b8\u30c8\u30ea\u3092\u4fee\u6b63\u3057\u3066\u30c6\u30b9\u30c8\u306b\u4e0d\u5408\u683c\u306b\u3057\u3001\u305d\u306e\u30e2\u30c7\u30eb\u304c\u30b3\u30fc\u30c9\u30d9\u30fc\u30b9\u3092\u4fee\u6b63\u3057\u3066\u518d\u3073\u4e0d\u5408\u683c\u306e\u30c6\u30b9\u30c8\u306b\u5408\u683c\u3067\u304d\u308b\u304b\u3069\u3046\u304b\u3092\u78ba\u8a8d\u3057\u307e\u3059\u3002<\/p>\n<p>\u7570\u306a\u308b\u30d7\u30ed\u30b0\u30e9\u30df\u30f3\u30b0\u8a00\u8a9e\u3084\u30d5\u30ec\u30fc\u30e0\u30ef\u30fc\u30af\u3067\u3001\u3067\u304d\u308b\u9650\u308a\u591a\u304f\u306e\u7570\u306a\u308b\u30b7\u30ca\u30ea\u30aa\u3092\u4f5c\u6210\u3057\u307e\u3059\u3002\u307e\u305f\u3001\u30b5\u30dd\u30fc\u30c8\u3059\u308b\u8a00\u8a9e\u306e\u8907\u6570\u306e\u7570\u306a\u308b\u30d0\u30fc\u30b8\u30e7\u30f3\u3092\u4f7f\u7528\u3059\u308b\u306a\u3069\u3001\u5b9f\u884c\u3059\u308b\u30c6\u30b9\u30c8\u306e\u6570\u3092\u5e38\u306b\u5897\u3084\u3057\u3066\u3044\u307e\u3059\u3002\u3053\u306e\u4f5c\u696d\u306b\u306f\u591a\u304f\u306e\u6642\u9593\u304c\u304b\u304b\u308a\u307e\u3059\u304c\u3001\u30e2\u30c7\u30eb\u306e\u54c1\u8cea\u3092\u8a55\u4fa1\u3059\u308b\u6700\u5584\u306e\u65b9\u6cd5\u3067\u3059\u3002<\/p>\n<aside class=\"p-4 p-md-6 post-aside--large\" style=\"background-color:#f6f8fa; margin:30px 0px;\">\n<p style=\"font-size:20px; font-weight:bold;\">\u30aa\u30d5\u30e9\u30a4\u30f3\u8a55\u4fa1\u3067\u6e2c\u5b9a\u3059\u308b\u3082\u306e<\/p>\n<p>\u30b3\u30fc\u30c9\u306e\u5b8c\u6210\u5ea6<\/p>\n<ul>\n<li><strong>\u5408\u683c\u3057\u305f\u30e6\u30cb\u30c3\u30c8\u30c6\u30b9\u30c8\u306e\u5272\u5408\u3002<\/strong>\u610f\u56f3\u7684\u306b\u58ca\u308c\u305f\u30b3\u30fc\u30c9\u3092\u53ef\u80fd\u306a\u9650\u308a\u4fee\u6b63\u3067\u304d\u308b\u3050\u3089\u3044\u306f\u30e2\u30c7\u30eb\u304c\u5341\u5206\u306b\u512a\u308c\u3066\u3044\u308b\u3053\u3068\u3092\u671b\u3093\u3067\u3044\u307e\u3059\u3002<\/li>\n<li><strong>\u5143\u306e\u65e2\u77e5\u306e\u5408\u683c\u72b6\u614b\u3068\u306e\u985e\u4f3c\u6027\u3002<\/strong>\u30b3\u30fc\u30c9\u306e\u63d0\u6848\u306e\u8cea\u3092\u6e2c\u308b\u305f\u3081\u306b\u3053\u308c\u3092\u4f7f\u3044\u307e\u3059\u3002\u3082\u3061\u308d\u3093\u3001\u79c1\u305f\u3061\u306e\u30aa\u30ea\u30b8\u30ca\u30eb\u30d0\u30fc\u30b8\u30e7\u30f3\u3088\u308a\u3082\u826f\u3044\u30b3\u30fc\u30c9\u306e\u66f8\u304d\u65b9\u304c\u3042\u308b\u304b\u3082\u3057\u308c\u307e\u305b\u3093\u304c\u3001\u79c1\u305f\u3061\u306f\u30c6\u30b9\u30c8\u3059\u308b\u305f\u3081\u306e\u5f37\u56fa\u306a\u30d9\u30fc\u30b9\u30e9\u30a4\u30f3\u3092\u63d0\u4f9b\u3057\u307e\u3059\u3002<\/li>\n<\/ul>\n<p>GitHub Copilot Chat\u306e\u5834\u5408<\/p>\n<ul>\n<li><strong>\u6b63\u3057\u304f\u56de\u7b54\u3055\u308c\u305f\u8cea\u554f\u306e\u5272\u5408\u3002<\/strong>\u30c1\u30e3\u30c3\u30c8\u304c\u63d0\u4f9b\u3059\u308b\u6280\u8853\u7684\u306a\u8cea\u554f\u306b\u5bfe\u3059\u308b\u56de\u7b54\u304c\u53ef\u80fd\u306a\u9650\u308a\u6b63\u78ba\u3067\u3042\u308b\u3053\u3068\u3092\u78ba\u8a8d\u3057\u307e\u3059\u3002<\/li>\n<\/ul>\n<p>\u4e21\u65b9\u306b\u3064\u3044\u3066\uff1a<\/p>\n<ul>\n<li><strong>\u30c8\u30fc\u30af\u30f3\u306e\u4f7f\u7528\u91cf\u3002<\/strong>\u3053\u308c\u306f\u4e3b\u8981\u306a\u30e2\u30c7\u30eb\u306e\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u6307\u6a19\u306e1\u3064\u3067\u3059\u3002\u901a\u5e38\u3001\u3042\u308b\u7d50\u679c\u3092\u5f97\u308b\u305f\u3081\u306b\u5fc5\u8981\u306a\u30c8\u30fc\u30af\u30f3\u306e\u6570\u304c\u5c11\u306a\u3044\u30e2\u30c7\u30eb\u307b\u3069\u52b9\u7387\u7684\u3067\u3059\u3002<\/li>\n<\/ul>\n<\/aside>\n<h2 id=\"using-ai-to-test-ai\">AI\u3092\u30c6\u30b9\u30c8\u3059\u308b\u305f\u3081\u306bAI\u3092\u4f7f\u7528\u3059\u308b<\/h2>\n<p>GitHub Copilot\u306f\u30b3\u30fc\u30c9\u3092\u751f\u6210\u3059\u308b\u3060\u3051\u3067\u306f\u3042\u308a\u307e\u305b\u3093\u3002GitHub Copilot Chat\u3067\u306f\u3001\u30b3\u30fc\u30c9\u306b\u95a2\u3059\u308b\u8cea\u554f\u306b\u7b54\u3048\u305f\u308a\u3001\u554f\u984c\u89e3\u6c7a\u3078\u306e\u3055\u307e\u3056\u307e\u306a\u30a2\u30d7\u30ed\u30fc\u30c1\u3092\u63d0\u6848\u3057\u305f\u308a\u3067\u304d\u307e\u3059\u3002\u79c1\u305f\u3061\u306f\u3001\u30e2\u30c7\u30eb\u306e\u30c1\u30e3\u30c3\u30c8\u6a5f\u80fd\u306e\u54c1\u8cea\u3092\u8a55\u4fa1\u3059\u308b\u305f\u3081\u306b\u4f7f\u7528\u3059\u308b1,000\u4ee5\u4e0a\u306e\u6280\u8853\u7684\u306a\u8cea\u554f\u3092\u96c6\u3081\u3066\u3044\u307e\u3059\u3002\u3053\u308c\u3089\u306e\u4e2d\u306b\u306f\u3001\u79c1\u305f\u3061\u304c\u7c21\u5358\u306b\u81ea\u52d5\u8a55\u4fa1\u3067\u304d\u308b\u5358\u7d14\u306a\u771f\u507d\u3092\u554f\u3046\u8cea\u554f\u3082\u3042\u308a\u307e\u3059\u3002\u3057\u304b\u3057\u3001\u3088\u308a\u8907\u96d1\u306a\u8cea\u554f\u306b\u3064\u3044\u3066\u306f\u3001\u5225\u306eLLM\u3092\u4f7f\u7528\u3057\u3066\u3001\u8a55\u4fa1\u5bfe\u8c61\u306e\u30e2\u30c7\u30eb\u304c\u63d0\u4f9b\u3059\u308b\u56de\u7b54\u3092\u30c1\u30a7\u30c3\u30af\u3057\u307e\u3059\u3002<\/p>\n<p>\u3053\u306e\u3088\u3046\u306a\u76ee\u7684\u306b\u306f\u3001\u79c1\u305f\u3061\u306e\u4f5c\u696d\u5168\u4f53\u306b\u308f\u305f\u3063\u3066\u4e00\u8cab\u3057\u305f\u8a55\u4fa1\u3092\u4fdd\u8a3c\u3059\u308b\u305f\u3081\u306b\u3001\u512a\u308c\u305f\u30d1\u30d5\u30a9\u30fc\u30de\u30f3\u30b9\u3092\u6301\u3064\u3053\u3068\u304c\u5206\u304b\u3063\u3066\u3044\u308b\u30e2\u30c7\u30eb\u3092\u4f7f\u7528\u3057\u307e\u3059\u3002\u307e\u305f\u3001\u3053\u306eLLM\u304c\u6b63\u3057\u304f\u52d5\u4f5c\u3057\u3066\u3044\u308b\u3053\u3068\u3092\u78ba\u8a8d\u3059\u308b\u305f\u3081\u306b\u3001\u8a55\u4fa1\u30b7\u30ca\u30ea\u30aa\u306e\u4e2d\u3067\u3053\u306eLLM\u306e\u51fa\u529b\u3092\u65e5\u5e38\u7684\u306b\u76e3\u67fb\u3057\u3066\u3044\u307e\u3059\u3002\u56de\u7b54\u3092\u8a55\u4fa1\u3059\u308b\u305f\u3081\u306b\u4f7f\u7528\u3059\u308bLLM\u304c\u3001\u4eba\u9593\u306e\u30ec\u30d3\u30e5\u30a2\u30fc\u3068\u4e00\u81f4\u3057\u3001\u591a\u304f\u306e\u30ea\u30af\u30a8\u30b9\u30c8\u306b\u308f\u305f\u3063\u3066\u4e00\u8cab\u3057\u305f\u6027\u80fd\u3092\u767a\u63ee\u3059\u308b\u3053\u3068\u3092\u4fdd\u8a3c\u3059\u308b\u306e\u306f\u5927\u5909\u306a\u3053\u3068\u3067\u3059\u3002<\/p>\n<p>\u79c1\u305f\u3061\u306f\u307e\u305f\u3001\u3053\u308c\u3089\u306e\u30c6\u30b9\u30c8\u3092\u6bce\u65e5\u672c\u756a\u30e2\u30c7\u30eb\u306b\u5bfe\u3057\u3066\u5b9f\u884c\u3057\u3066\u3044\u307e\u3059\u3002\u52a3\u5316\u304c\u898b\u3089\u308c\u305f\u5834\u5408\u306f\u3001\u76e3\u67fb\u3092\u884c\u3063\u3066\u3001\u30e2\u30c7\u30eb\u304c\u4ee5\u524d\u307b\u3069\u3046\u307e\u304f\u6a5f\u80fd\u3057\u3066\u3044\u306a\u3044\u7406\u7531\u3092\u63a2\u308a\u307e\u3059\u3002\u6642\u306b\u306f\u3001\u79c1\u305f\u3061\u304c\u671f\u5f85\u3059\u308b\u54c1\u8cea\u30ec\u30d9\u30eb\u306b\u623b\u3059\u305f\u3081\u306b\u3001\u30d7\u30ed\u30f3\u30d7\u30c8\u306e\u4e00\u90e8\u3092\u4fee\u6b63\u3059\u308b\u306a\u3069\u3001\u5909\u66f4\u3092\u52a0\u3048\u308b\u5fc5\u8981\u304c\u3042\u308a\u307e\u3059\u3002<\/p>\n<h2 id=\"running-the-tests\">\u30c6\u30b9\u30c8\u306e\u5b9f\u884c<\/h2>\n<p>\u79c1\u305f\u3061\u306e\u30bb\u30c3\u30c8\u30a2\u30c3\u30d7\u304c\u512a\u308c\u3066\u3044\u308b\u70b9\u306f\u3001\u88fd\u54c1\u30b3\u30fc\u30c9\u3092\u5909\u66f4\u3059\u308b\u3053\u3068\u306a\u304f\u65b0\u3057\u3044\u30e2\u30c7\u30eb\u3092\u30c6\u30b9\u30c8\u3067\u304d\u308b\u3053\u3068\u3067\u3059\u3002\u79c1\u305f\u3061\u306e\u30a4\u30f3\u30d5\u30e9\u306b\u306f\u3001\u30b3\u30fc\u30c9\u88dc\u5b8c\u6a5f\u80fd\u304c\u4f7f\u7528\u3059\u308b\u30d7\u30ed\u30ad\u30b7\u30b5\u30fc\u30d0\u30fc\u304c\u7d44\u307f\u8fbc\u307e\u308c\u3066\u3044\u307e\u3059\u3002\u30af\u30e9\u30a4\u30a2\u30f3\u30c8\u5074\u3092\u5909\u66f4\u3059\u308b\u3053\u3068\u306a\u304f\u3001\u30d7\u30ed\u30ad\u30b7\u30b5\u30fc\u30d0\u30fc\u304c\u30ec\u30b9\u30dd\u30f3\u30b9\u306e\u305f\u3081\u306b\u547c\u3073\u51fa\u3059API\u3092\u7c21\u5358\u306b\u5909\u66f4\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002\u3053\u308c\u306b\u3088\u308a\u3001\u88fd\u54c1\u30b3\u30fc\u30c9\u3092\u5909\u66f4\u3059\u308b\u3053\u3068\u306a\u304f\u3001\u65b0\u3057\u3044\u30e2\u30c7\u30eb\u3092\u8fc5\u901f\u306b\u53cd\u5fa9\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002<\/p>\n<p>\u3053\u308c\u3089\u306e\u30c6\u30b9\u30c8\u306f\u3059\u3079\u3066\u3001\u4e3b\u306bGitHub Actions\u3067\u69cb\u7bc9\u3055\u308c\u305f\u72ec\u81ea\u306e\u30ab\u30b9\u30bf\u30e0\u30d7\u30e9\u30c3\u30c8\u30d5\u30a9\u30fc\u30e0\u4e0a\u3067\u5b9f\u884c\u3055\u308c\u307e\u3059\u3002\u7d50\u679c\u306f\u3001Apache Kafka\u3084Microsoft Azure\u306e\u3088\u3046\u306a\u30b7\u30b9\u30c6\u30e0\u304b\u3089\u51fa\u3057\u5165\u308c\u3055\u308c\u3001\u30c7\u30fc\u30bf\u3092\u63a2\u7d22\u3059\u308b\u305f\u3081\u306b\u69d8\u3005\u306a\u30c0\u30c3\u30b7\u30e5\u30dc\u30fc\u30c9\u3092\u6d3b\u7528\u3057\u3066\u3044\u307e\u3059\u3002<\/p>\n<h2 id=\"making-the-call-to-adopt-or-not\">\u63a1\u7528\u3059\u308b\u304b\u3069\u3046\u304b\u306e\u5224\u65ad<\/h2>\n<p>\u53ce\u96c6\u3057\u305f\u30c7\u30fc\u30bf\u3092\u3069\u3046\u3059\u308b\u304b\u304c\u5927\u304d\u306a\u8ab2\u984c\u3067\u3059\u3002\u3042\u308b\u30e2\u30c7\u30eb\u304c\u5168\u4f53\u7684\u306b\u4e0d\u632f\u3060\u3063\u305f\u5834\u5408\u306a\u3069\u3001\u6c7a\u65ad\u304c\u7c21\u5358\u306a\u3053\u3068\u3082\u3042\u308a\u307e\u3059\u3002\u3057\u304b\u3057\u3001\u3042\u308b\u30e2\u30c7\u30eb\u304c\u53d7\u5165\u7387\u3092\u5927\u5e45\u306b\u5411\u4e0a\u3055\u305b\u305f\u304c\u3001\u540c\u6642\u306b\u5f85\u3061\u6642\u9593\u3082\u5897\u52a0\u3055\u305b\u305f\u3068\u3057\u305f\u3089\u3069\u3046\u3067\u3057\u3087\u3046\u304b\uff1f<\/p>\n<p>\u30e1\u30c8\u30ea\u30af\u30b9\u306e\u9593\u306b\u306f\u6642\u306b\u9006\u76f8\u95a2\u304c\u3042\u308a\u307e\u3059\u3002\u5f85\u3061\u6642\u9593\u304c\u9577\u304f\u306a\u308b\u3068\u3001\u30e6\u30fc\u30b6\u30fc\u304c\u898b\u308b\u63d0\u6848\u306e\u6570\u304c\u5c11\u306a\u304f\u306a\u308b\u305f\u3081\u3001\u5b9f\u969b\u306b\u306f\u53d7\u7406\u7387\u304c\u9ad8\u304f\u306a\u308b\u304b\u3082\u3057\u308c\u307e\u305b\u3093\u3002<\/p>\n<p>GitHub\u306e\u76ee\u6a19\u306f\u3001\u53ef\u80fd\u306a\u9650\u308a\u6700\u9ad8\u54c1\u8cea\u3067\u8cac\u4efb\u3042\u308bAI\u30b3\u30fc\u30c7\u30a3\u30f3\u30b0\u30fb\u30a2\u30b7\u30b9\u30bf\u30f3\u30c8\u3092\u4f5c\u308b\u3053\u3068\u3067\u3042\u308a\u3001\u305d\u308c\u304c\u3001\u88fd\u54c1\u5185\u3067\u3069\u306e\u30e2\u30c7\u30eb\u3092\u30b5\u30dd\u30fc\u30c8\u3059\u308b\u304b\u3068\u3044\u3046\u6c7a\u65ad\u306e\u6307\u91dd\u3068\u306a\u3063\u3066\u3044\u307e\u3059\u3002\u3057\u304b\u3057\u3001\u79c1\u305f\u3061\u306e\u8a55\u4fa1\u624b\u9806\u304c\u63d0\u4f9b\u3059\u308b\u8cea\u306e\u9ad8\u3044\u30c7\u30fc\u30bf\u306a\u3057\u306b\u306f\u3001\u305d\u308c\u3092\u4f5c\u308b\u3053\u3068\u306f\u3067\u304d\u307e\u305b\u3093\u3002\u3053\u306e\u8a18\u4e8b\u304c\u3001\u3042\u306a\u305f\u81ea\u8eab\u306e\u30e6\u30fc\u30b9\u30b1\u30fc\u30b9\u306b\u9069\u7528\u3067\u304d\u308b\u30a2\u30a4\u30c7\u30a2\u3092\u63d0\u4f9b\u3067\u304d\u3066\u3044\u308c\u3070\u5e78\u3044\u3067\u3059\u3002<\/p>\n<hr style=\"margin-top: 50px;\" \/>\n<div class=\"post-content-cta\">\n<strong>GitHub Models\u3092\u4f7f\u3063\u3066\u751f\u6210AI\u30a2\u30d7\u30ea\u30b1\u30fc\u30b7\u30e7\u30f3\u3092\u4f5c\u308d\u3046<\/strong><\/p>\n<p>GitHub Models\u3092\u4f7f\u3048\u3070\u3001OpenAI\u3001Cohere\u3001Microsoft\u3001Mistral\u306a\u3069\u306eAI\u30e2\u30c7\u30eb\u3092\u7c21\u5358\u306b\u5229\u7528\u3001\u6bd4\u8f03\u3001\u5b9f\u9a13\u3001\u69cb\u7bc9\u3059\u308b\u3053\u3068\u304c\u3067\u304d\u307e\u3059\u3002<\/p>\n<p><a href=\"https:\/\/github.com\/marketplace\/models?utm_source=blog&amp;utm_medium=how_github_evaluates_ai_models&amp;utm_campaign=selenium\">\u4eca\u3059\u3050\u304a\u8a66\u3057\u304f\u3060\u3055\u3044\u3002<\/a><\/p>\n<\/div>\n<hr style=\"margin-bottom: 50px;\" \/>\n<\/div>\n<p>The post<a href=\"https:\/\/github.blog\/ai-and-ml\/generative-ai\/how-we-evaluate-models-for-github-copilot\/\"> How we evaluate models for GitHub Copilot <\/a>appeared first on<a href=\"https:\/\/github.blog\"> The GitHub Blog<\/a>.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u672c\u8a18\u4e8b\u3067\u306f\u3001GitHub Copilot \u30c1\u30fc\u30e0\u304c AI \u30e2\u30c7\u30eb\u3092\u8a55\u4fa1\u3059\u308b\u969b\u306e\u7d4c\u9a13\u306e\u4e00\u90e8\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002\u7279\u306b\u3001\u672c\u756a\u74b0\u5883\u306b\u5909\u66f4\u3092\u52a0\u3048\u308b\u524d\u306b\u5b9f\u65bd\u3059\u308b\u30aa\u30d5\u30e9\u30a4\u30f3\u8a55\u4fa1\u306b\u7126\u70b9\u3092\u5f53\u3066\u3066\u3044\u307e\u3059\u3002<\/p>\n","protected":false},"author":2109,"featured_media":5066,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"jetpack_post_was_ever_published":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2},"_wpas_customize_per_network":false},"categories":[152],"tags":[277,173],"coauthors":[162],"class_list":["post-5065","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-github-copilot","tag-ai-ml","tag-generative-ai"],"yoast_head":"<!-- This site is optimized with the Yoast SEO Premium plugin v27.3 (Yoast SEO v27.3) - https:\/\/yoast.com\/product\/yoast-seo-premium-wordpress\/ -->\n<title>GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5 - GitHub\u30d6\u30ed\u30b0<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5\" \/>\n<meta property=\"og:description\" content=\"\u672c\u8a18\u4e8b\u3067\u306f\u3001GitHub Copilot \u30c1\u30fc\u30e0\u304c AI \u30e2\u30c7\u30eb\u3092\u8a55\u4fa1\u3059\u308b\u969b\u306e\u7d4c\u9a13\u306e\u4e00\u90e8\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002\u7279\u306b\u3001\u672c\u756a\u74b0\u5883\u306b\u5909\u66f4\u3092\u52a0\u3048\u308b\u524d\u306b\u5b9f\u65bd\u3059\u308b\u30aa\u30d5\u30e9\u30a4\u30f3\u8a55\u4fa1\u306b\u7126\u70b9\u3092\u5f53\u3066\u3066\u3044\u307e\u3059\u3002\" \/>\n<meta property=\"og:url\" content=\"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/\" \/>\n<meta property=\"og:site_name\" content=\"GitHub\u30d6\u30ed\u30b0\" \/>\n<meta property=\"article:published_time\" content=\"2025-03-06T09:00:57+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/github.blog\/jp\/wp-content\/uploads\/sites\/2\/2025\/03\/copilot-header-2.png?fit=1600%2C840\" \/>\n\t<meta property=\"og:image:width\" content=\"1600\" \/>\n\t<meta property=\"og:image:height\" content=\"840\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"author\" content=\"tomokota\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"tomokota\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"1 minute\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/\"},\"author\":{\"name\":\"tomokota\",\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/#\\\/schema\\\/person\\\/7801e74a234e0a6d7922ab479a6218dc\"},\"headline\":\"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5\",\"datePublished\":\"2025-03-06T09:00:57+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/\"},\"wordCount\":119,\"image\":{\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/github.blog\\\/jp\\\/wp-content\\\/uploads\\\/sites\\\/2\\\/2025\\\/03\\\/copilot-header-2.png?fit=1600%2C840\",\"keywords\":[\"AI &amp; ML\",\"generative AI\"],\"articleSection\":[\"GitHub Copilot\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/\",\"url\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/\",\"name\":\"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5 - GitHub\u30d6\u30ed\u30b0\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/github.blog\\\/jp\\\/wp-content\\\/uploads\\\/sites\\\/2\\\/2025\\\/03\\\/copilot-header-2.png?fit=1600%2C840\",\"datePublished\":\"2025-03-06T09:00:57+00:00\",\"author\":{\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/#\\\/schema\\\/person\\\/7801e74a234e0a6d7922ab479a6218dc\"},\"breadcrumb\":{\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/#primaryimage\",\"url\":\"https:\\\/\\\/github.blog\\\/jp\\\/wp-content\\\/uploads\\\/sites\\\/2\\\/2025\\\/03\\\/copilot-header-2.png?fit=1600%2C840\",\"contentUrl\":\"https:\\\/\\\/github.blog\\\/jp\\\/wp-content\\\/uploads\\\/sites\\\/2\\\/2025\\\/03\\\/copilot-header-2.png?fit=1600%2C840\",\"width\":1600,\"height\":840},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/github.blog\\\/jp\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/#website\",\"url\":\"https:\\\/\\\/github.blog\\\/jp\\\/\",\"name\":\"GitHub\u30d6\u30ed\u30b0\",\"description\":\"\u88fd\u54c1\u30a2\u30c3\u30d7\u30c7\u30fc\u30c8\u3084\u958b\u767a\u306b\u95a2\u3059\u308b\u30a2\u30a4\u30c7\u30a3\u30a2\u3084\u30a4\u30f3\u30b9\u30d4\u30ec\u30fc\u30b7\u30e7\u30f3\u306a\u3069\u3001\u30a8\u30f3\u30b8\u30cb\u30a2\u306e\u7686\u3055\u3093\u306b\u5f79\u7acb\u3064\u60c5\u5831\u3092\u767a\u4fe1\u3057\u307e\u3059\u3002\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/github.blog\\\/jp\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/github.blog\\\/jp\\\/#\\\/schema\\\/person\\\/7801e74a234e0a6d7922ab479a6218dc\",\"name\":\"tomokota\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/956ea2b792fa8577ae271729cccae2d338be2fb17ff928ff517892ab125d3d04?s=96&d=mm&r=g1cd54adb51b368ddd9aab5c3580bf6a4\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/956ea2b792fa8577ae271729cccae2d338be2fb17ff928ff517892ab125d3d04?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/956ea2b792fa8577ae271729cccae2d338be2fb17ff928ff517892ab125d3d04?s=96&d=mm&r=g\",\"caption\":\"tomokota\"},\"sameAs\":[\"tomokota\"],\"url\":\"https:\\\/\\\/github.blog\\\/jp\\\/author\\\/tomokota\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO Premium plugin. -->","yoast_head_json":{"title":"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5 - GitHub\u30d6\u30ed\u30b0","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/","og_locale":"en_US","og_type":"article","og_title":"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5","og_description":"\u672c\u8a18\u4e8b\u3067\u306f\u3001GitHub Copilot \u30c1\u30fc\u30e0\u304c AI \u30e2\u30c7\u30eb\u3092\u8a55\u4fa1\u3059\u308b\u969b\u306e\u7d4c\u9a13\u306e\u4e00\u90e8\u3092\u7d39\u4ecb\u3057\u307e\u3059\u3002\u7279\u306b\u3001\u672c\u756a\u74b0\u5883\u306b\u5909\u66f4\u3092\u52a0\u3048\u308b\u524d\u306b\u5b9f\u65bd\u3059\u308b\u30aa\u30d5\u30e9\u30a4\u30f3\u8a55\u4fa1\u306b\u7126\u70b9\u3092\u5f53\u3066\u3066\u3044\u307e\u3059\u3002","og_url":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/","og_site_name":"GitHub\u30d6\u30ed\u30b0","article_published_time":"2025-03-06T09:00:57+00:00","og_image":[{"width":1600,"height":840,"url":"https:\/\/github.blog\/jp\/wp-content\/uploads\/sites\/2\/2025\/03\/copilot-header-2.png?fit=1600%2C840","type":"image\/png"}],"author":"tomokota","twitter_card":"summary_large_image","twitter_misc":{"Written by":"tomokota","Est. reading time":"1 minute"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/#article","isPartOf":{"@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/"},"author":{"name":"tomokota","@id":"https:\/\/github.blog\/jp\/#\/schema\/person\/7801e74a234e0a6d7922ab479a6218dc"},"headline":"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5","datePublished":"2025-03-06T09:00:57+00:00","mainEntityOfPage":{"@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/"},"wordCount":119,"image":{"@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/#primaryimage"},"thumbnailUrl":"https:\/\/github.blog\/jp\/wp-content\/uploads\/sites\/2\/2025\/03\/copilot-header-2.png?fit=1600%2C840","keywords":["AI &amp; ML","generative AI"],"articleSection":["GitHub Copilot"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/","url":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/","name":"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5 - GitHub\u30d6\u30ed\u30b0","isPartOf":{"@id":"https:\/\/github.blog\/jp\/#website"},"primaryImageOfPage":{"@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/#primaryimage"},"image":{"@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/#primaryimage"},"thumbnailUrl":"https:\/\/github.blog\/jp\/wp-content\/uploads\/sites\/2\/2025\/03\/copilot-header-2.png?fit=1600%2C840","datePublished":"2025-03-06T09:00:57+00:00","author":{"@id":"https:\/\/github.blog\/jp\/#\/schema\/person\/7801e74a234e0a6d7922ab479a6218dc"},"breadcrumb":{"@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/#primaryimage","url":"https:\/\/github.blog\/jp\/wp-content\/uploads\/sites\/2\/2025\/03\/copilot-header-2.png?fit=1600%2C840","contentUrl":"https:\/\/github.blog\/jp\/wp-content\/uploads\/sites\/2\/2025\/03\/copilot-header-2.png?fit=1600%2C840","width":1600,"height":840},{"@type":"BreadcrumbList","@id":"https:\/\/github.blog\/jp\/2025-03-06-ai-and-ml-generative-ai-how-we-evaluate-models-for-github-copilot\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/github.blog\/jp\/"},{"@type":"ListItem","position":2,"name":"GitHub Copilot\u306eAI\u30e2\u30c7\u30eb\u3068LLM\u306e\u8a55\u4fa1\u65b9\u6cd5"}]},{"@type":"WebSite","@id":"https:\/\/github.blog\/jp\/#website","url":"https:\/\/github.blog\/jp\/","name":"GitHub\u30d6\u30ed\u30b0","description":"\u88fd\u54c1\u30a2\u30c3\u30d7\u30c7\u30fc\u30c8\u3084\u958b\u767a\u306b\u95a2\u3059\u308b\u30a2\u30a4\u30c7\u30a3\u30a2\u3084\u30a4\u30f3\u30b9\u30d4\u30ec\u30fc\u30b7\u30e7\u30f3\u306a\u3069\u3001\u30a8\u30f3\u30b8\u30cb\u30a2\u306e\u7686\u3055\u3093\u306b\u5f79\u7acb\u3064\u60c5\u5831\u3092\u767a\u4fe1\u3057\u307e\u3059\u3002","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/github.blog\/jp\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/github.blog\/jp\/#\/schema\/person\/7801e74a234e0a6d7922ab479a6218dc","name":"tomokota","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/956ea2b792fa8577ae271729cccae2d338be2fb17ff928ff517892ab125d3d04?s=96&d=mm&r=g1cd54adb51b368ddd9aab5c3580bf6a4","url":"https:\/\/secure.gravatar.com\/avatar\/956ea2b792fa8577ae271729cccae2d338be2fb17ff928ff517892ab125d3d04?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/956ea2b792fa8577ae271729cccae2d338be2fb17ff928ff517892ab125d3d04?s=96&d=mm&r=g","caption":"tomokota"},"sameAs":["tomokota"],"url":"https:\/\/github.blog\/jp\/author\/tomokota\/"}]}},"jetpack_publicize_connections":[],"jetpack_featured_media_url":"https:\/\/github.blog\/jp\/wp-content\/uploads\/sites\/2\/2025\/03\/copilot-header-2.png?fit=1600%2C840","jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/pasZHI-1jH","_links":{"self":[{"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/posts\/5065","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/users\/2109"}],"replies":[{"embeddable":true,"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/comments?post=5065"}],"version-history":[{"count":6,"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/posts\/5065\/revisions"}],"predecessor-version":[{"id":5072,"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/posts\/5065\/revisions\/5072"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/media\/5066"}],"wp:attachment":[{"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/media?parent=5065"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/categories?post=5065"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/tags?post=5065"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/github.blog\/jp\/wp-json\/wp\/v2\/coauthors?post=5065"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}