{"id":725730,"date":"2026-07-23T04:44:42","date_gmt":"2026-07-23T03:44:42","guid":{"rendered":"https:\/\/blog.jetbrains.com\/?post_type=kotlin&#038;p=725730"},"modified":"2026-07-23T04:44:45","modified_gmt":"2026-07-23T03:44:45","slug":"introducing-the-kotlin-benchmark-evaluate-ai-coding-agents-on-real-world-kotlin-tasks","status":"publish","type":"kotlin","link":"https:\/\/blog.jetbrains.com\/zh-hans\/kotlin\/2026\/07\/introducing-the-kotlin-benchmark-evaluate-ai-coding-agents-on-real-world-kotlin-tasks\/","title":{"rendered":"\u4e3a AI \u7f16\u7801\u667a\u80fd\u4f53\u63a8\u51fa Kotlin Benchmark"},"content":{"rendered":"<p>\u667a\u80fd\u4f53\u5316\u7f16\u7801\u57fa\u51c6\u6b63\u9010\u6b65\u8d34\u8fd1\u771f\u5b9e\u8f6f\u4ef6\u5f00\u53d1\u3002\u5bf9\u4e8e Kotlin \u56e2\u961f\u800c\u8a00\uff0c\u6700\u91cd\u8981\u7684\u95ee\u9898\u662f AI \u667a\u80fd\u4f53\u5b8c\u6210\u7aef\u5230\u7aef Kotlin \u4efb\u52a1\uff08\u4ece\u8bfb\u53d6\u95ee\u9898\u5230\u751f\u6210\u53ef\u4ee5\u901a\u8fc7\u9a8c\u8bc1\u7684\u89e3\u51b3\u65b9\u6848\uff09\u7684\u53ef\u9760\u6027\u5982\u4f55\u3002<\/p>\n<p>\u4f5c\u4e3a\u89e3\u51b3\u8fd9\u4e00\u95ee\u9898\u7684\u7b2c\u4e00\u6b65\uff0c\u6211\u4eec\u53d1\u5e03\u4e86 <a href=\"http:\/\/kotlinlang.org\/benchmark\/\" target=\"_blank\" rel=\"noopener\">Kotlin Benchmark<\/a>\uff0cKotlin Benchmark \u662f JetBrains \u7528\u4e8e\u8bc4\u4f30 AI \u7f16\u7801\u667a\u80fd\u4f53\u5728\u5904\u7406 Kotlin \u8f6f\u4ef6\u5de5\u7a0b\u4efb\u52a1\u65f6\u8868\u73b0\u7684\u5b98\u65b9\u57fa\u51c6\u3002\u6211\u4eec\u7684\u76ee\u6807\u662f\u4e3a\u5f00\u53d1\u8005\u63d0\u4f9b\u4e00\u79cd\u53ef\u4fe1\u3001\u516c\u5f00\u7684\u65b9\u5f0f\u6765\u8bc4\u4f30\u4e0d\u540c\u667a\u80fd\u4f53\u5728 Kotlin \u4e0a\u7684\u8868\u73b0\uff0c\u5e76\u4f7f\u7528\u66f4\u8d34\u8fd1\u65e5\u5e38\u5f00\u53d1\u5de5\u4f5c\u7684\u4efb\u52a1\u6765\u6bd4\u8f83\u4e0d\u540c\u7684\u667a\u80fd\u4f53\u8bbe\u7f6e\u3002<\/p>\n<p>\u4f34\u968f\u57fa\u51c6\u7684\u53d1\u5e03\uff0c\u6211\u4eec\u4e5f\u5c06\u5728 GitHub \u4e0a\u53d1\u5e03\u57fa\u51c6\u8d44\u6e90\uff0c\u5e76\u63a8\u51fa\u5b98\u65b9\u6392\u884c\u699c\u6765\u8ddf\u8e2a\u8bc4\u4f30\u7ed3\u679c\u3002<\/p>\n<p><a href=\"https:\/\/github.com\/Kotlin\/kotlin-swe-bench\" target=\"_blank\" rel=\"noopener\">\u5728 GitHub \u4e0a\u63a2\u7d22\u57fa\u51c6<\/a><\/p>\n<p><a href=\"http:\/\/kotlinlang.org\/benchmark\/\" target=\"_blank\" rel=\"noopener\">\u67e5\u770b\u6392\u884c\u699c\u4e0a\u7684\u9996\u6279\u7ed3\u679c<\/a><\/p>\n<h3 class=\"wp-block-heading\">Kotlin Benchmark \u7684\u8fd0\u4f5c\u65b9\u5f0f<\/h3>\n<p>Kotlin Benchmark \u7684\u9996\u4e2a\u516c\u5f00\u7248\u672c\u57fa\u4e8e SWE-bench \u65b9\u6cd5\uff0c\u4e14\u4fa7\u91cd\u4e8e\u4ed3\u5e93\u7ea7 Kotlin \u8f6f\u4ef6\u5de5\u7a0b\u4efb\u52a1\u3002<\/p>\n<p>Kotlin \u5df2\u62e5\u6709\u5f3a\u5927\u7684\u4fa7\u91cd\u4e8e\u6a21\u578b\u7684\u8bc4\u4f30\u8d44\u6e90\uff0c\u5305\u62ec <a href=\"https:\/\/huggingface.co\/datasets\/JetBrains\/Kotlin_HumanEval\" target=\"_blank\" rel=\"noopener\">Kotlin_HumanEval<\/a> \u548c <a href=\"https:\/\/huggingface.co\/datasets\/JetBrains\/Kotlin_QA\" target=\"_blank\" rel=\"noopener\">Kotlin_QA<\/a>\uff0c\u8fd9\u4e9b\u8bc4\u4f30\u8d44\u6e90\u53ef\u4ee5\u5e2e\u52a9\u8861\u91cf\u6a21\u578b\u5bf9\u8bed\u8a00\u7684\u8bed\u6cd5\u4e0e\u6838\u5fc3\u6982\u5ff5\u7684\u7406\u89e3\u3002\u800c Kotlin Benchmark \u5219\u805a\u7126\u4e8e\u53e6\u4e00\u5c42\u9762\uff1aAI \u7f16\u7801\u667a\u80fd\u4f53\u5728\u73b0\u6709 Kotlin \u9879\u76ee\u4e2d\u5b8c\u6210\u5df2\u9a8c\u8bc1\u8f6f\u4ef6\u5de5\u7a0b\u4efb\u52a1\u7684\u80fd\u529b\u3002<\/p>\n<p>\u6570\u636e\u96c6\u5305\u542b 105 \u4e2a\u6e90\u81ea\u6d3b\u8dc3\u5f00\u6e90\u4ed3\u5e93\u7684\u5de5\u7a0b\u4efb\u52a1\u3002\u6bcf\u9879\u4efb\u52a1\u90fd\u8981\u6c42 AI \u667a\u80fd\u4f53\u89e3\u8bfb\u771f\u5b9e\u95ee\u9898\u63cf\u8ff0\u3001\u68b3\u7406\u9879\u76ee\u4e0a\u4e0b\u6587\uff0c\u4ee5\u53ca\u751f\u6210\u53ef\u7528\u8865\u4e01\u3002\u89e3\u51b3\u65b9\u6848\u4f1a\u5728\u5bb9\u5668\u5316\u73af\u5883\u4e2d\u8fdb\u884c\u4e25\u683c\u9a8c\u8bc1\uff0c\u53ea\u6709\u751f\u6210\u7684\u89e3\u51b3\u65b9\u6848\u901a\u8fc7\u8981\u6c42\u7684\u6d4b\u8bd5\u9a8c\u8bc1\uff0c\u8be5\u4efb\u52a1\u624d\u4f1a\u88ab\u6807\u8bb0\u4e3a\u5df2\u89e3\u51b3\u3002<\/p>\n<p>\u8981\u8be6\u7ec6\u4e86\u89e3\u6211\u4eec\u7684\u73af\u5883\u8bbe\u7f6e\u548c\u6570\u636e\u91c7\u96c6\uff0c\u8bf7\u53c2\u9605<a href=\"https:\/\/kotlinlang.org\/benchmark\/methodology\/\" target=\"_blank\" rel=\"noopener\">\u65b9\u6cd5<\/a>\u9875\u9762\u3002<\/p>\n<h3 class=\"wp-block-heading\">\u9996\u6279\u7ed3\u679c<\/h3>\n<p>\u9996\u8f6e\u8bc4\u4f30\u8868\u660e\uff0c\u9886\u5148\u7684\u7f16\u7801\u667a\u80fd\u4f53\u53ef\u4ee5\u5b8c\u6210\u5f53\u524d Kotlin Benchmark \u4efb\u52a1\u4e2d\u7684\u5927\u90e8\u5206\u3002\u8fd9\u4e9b\u7ed3\u679c\u53cd\u6620\u4e86\u57fa\u51c6\u7684\u9996\u4e2a\u516c\u5f00\u7248\u672c\uff0c\u5c1a\u672a\u5305\u542b\u6700\u65b0\u53d1\u5e03\u7684\u6a21\u578b\u3002\u6211\u4eec\u5df2\u7740\u624b\u5f00\u53d1\u7b2c\u4e8c\u7248\u57fa\u51c6\uff0c\u5e76\u5c06\u968f\u7740\u65b0\u589e\u8bc4\u4f30\u7ed3\u679c\u7684\u52a0\u5165\u66f4\u65b0\u6392\u884c\u699c\u3002<\/p>\n<p>\u5728\u6b64\u6b21\u8fd0\u884c\u4e2d\uff0c\u6700\u597d\u7684\u7ed3\u679c\u6765\u81ea\u642d\u914d Opus 4.7 xhigh \u7684 Claude Code\uff0c\u5b83\u89e3\u51b3\u4e86 105 \u4e2a\u4efb\u52a1\u4e2d\u7684 90 \u4e2a\uff0c\u89e3\u51b3\u7387\u8fbe\u5230 85.71%\u3002\u642d\u914d Opus 4.7 max \u7684 JetBrains Junie\uff08\u89e3\u51b3\u7387 81.9%\uff09\u548c\u642d\u914d GPT 5.5 xhigh \u7684 Codex\uff08\u89e3\u51b3\u7387 81.9%\uff09\u7d27\u968f\u5176\u540e\u3002<\/p>\n<p>\u5b8c\u6574\u6392\u884c\u699c\u53ef\u4ee5\u5728 <a href=\"http:\/\/kotlinlang.org\/benchmark\" target=\"_blank\" rel=\"noopener\">kotlinlang.org\/benchmark<\/a> \u67e5\u770b\uff0c\u60a8\u53ef\u4ee5\u901a\u8fc7\u6b64\u6392\u884c\u699c\u8be6\u7ec6\u6bd4\u8f83\u4e0d\u540c\u667a\u80fd\u4f53\u548c\u914d\u7f6e\u7684\u8868\u73b0\u3002<\/p>\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" class=\"wp-image-720156\" src=\"https:\/\/blog.jetbrains.com\/wp-content\/uploads\/2026\/07\/SWE-Benchmark-Infographics-1920x1080-1.png\" alt=\"\" width=\"1920\" height=\"1080\" \/>\n<figcaption class=\"wp-element-caption\">\u6b64\u5904\u7ed9\u51fa\u7684\u7ed3\u679c\u53cd\u6620\u4e86 Kotlin Benchmark \u7684\u9996\u4e2a\u516c\u5f00\u7248\u672c\u7684\u8fd0\u884c\u7ed3\u679c\u3002\u6392\u884c\u699c\u5c06\u968f\u7740\u65b0\u589e\u6a21\u578b\u8bc4\u4f30\u7684\u52a0\u5165\u8fdb\u884c\u66f4\u65b0\u3002<\/figcaption>\n<\/figure>\n<p>\u5bf9\u4e8e\u8bc4\u4f30\u7f16\u7801\u667a\u80fd\u4f53\u7684\u56e2\u961f\uff0c\u8be5\u57fa\u51c6\u63d0\u4f9b\u4e86\u7edf\u4e00\u7684\u53c2\u8003\u6846\u67b6\u200c\uff0c\u7528\u4e8e\u6bd4\u8f83\u4e0d\u540c\u8bbe\u7f6e\u5728 Kotlin \u4efb\u52a1\u4e0a\u7684\u8868\u73b0\uff0c\u800c\u65e0\u9700\u4ec5\u4f9d\u8d56\u5382\u5546\u7684\u8bf4\u6cd5\u3002\u8fd9\u4e9b\u5206\u6570\u65e8\u5728\u4f5c\u4e3a\u53c2\u8003\uff0c\u800c\u975e\u5bf9\u6bcf\u4e00\u4e2a\u4ee3\u7801\u5e93\u7684\u4fdd\u8bc1\u3002\u5b9e\u9645\u7ed3\u679c\u53d6\u51b3\u4e8e\u60a8\u7684\u67b6\u6784\u3001\u5185\u90e8 API\u3001\u7f16\u7801\u6807\u51c6\u3001\u5de5\u5177\u548c\u9a8c\u8bc1\u6d41\u7a0b\u3002<\/p>\n<h3 class=\"wp-block-heading\">\u63a5\u4e0b\u6765\u6709\u4ec0\u4e48\u503c\u5f97\u671f\u5f85\u7684\uff1f<\/h3>\n<p>\u6211\u4eec\u91cd\u89c6\u5f00\u653e\u7684\u65b9\u5f0f\uff0c\u56e0\u6b64\uff0c\u6211\u4eec\u57fa\u4e8e\u5f00\u6e90 Multi-SWE-bench \u57fa\u7840\u67b6\u6784\u6784\u5efa\u4e86\u6b64\u57fa\u51c6\uff0c\u5e76\u516c\u5f00\u53d1\u5e03\u6240\u6709\u6570\u636e\u96c6\u548c\u6d4b\u8bd5\u5de5\u5177\u3002<\/p>\n<p>\u6211\u4eec\u5c06\u57fa\u51c6\u89c6\u4e3a\u6301\u7eed\u7684\u8d28\u91cf\u8bc4\u4f30\u6d41\u6c34\u7ebf\u3002\u672a\u6765\uff0c\u6211\u4eec\u8ba1\u5212\u5728\u4ee5\u4e0b\u51e0\u4e2a\u9886\u57df\u6269\u5c55\u8be5\u6846\u67b6\uff1a<\/p>\n<ul class=\"wp-block-list\">\n<li><strong>\u6269\u5c55 Kotlin \u751f\u6001\u7cfb\u7edf\u8986\u76d6\u8303\u56f4<\/strong>\uff1a\u6211\u4eec\u5e0c\u671b\u4efb\u52a1\u7ec4\u5408\u66f4\u8d34\u5408 Kotlin \u7684\u5b9e\u9645\u5e94\u7528\u573a\u666f\uff0c\u5305\u62ec Android\u3001Kotlin Multiplatform \u7b49\u9886\u57df\u7684\u4efb\u52a1\uff0c\u540c\u65f6\u8986\u76d6\u66f4\u5e7f\u6cdb\u7684\u4efb\u52a1\u96be\u5ea6\u7ea7\u522b\u3002<\/li>\n<li><strong>\u589e\u52a0\u8bc4\u4f30\u6307\u6807<\/strong>\uff1a\u901a\u8fc7\u6d4b\u8bd5\u53ea\u662f\u8861\u91cf\u6b63\u786e\u6027\u7684\u53c2\u8003\uff0c\u4f46\u8fd9\u53ea\u662f\u667a\u80fd\u4f53\u8bc4\u4f30\u7684\u5176\u4e2d\u4e00\u73af\u3002\u540e\u7eed\u7248\u672c\u5c06\u6ce8\u91cd\u6210\u672c\u3001\u6027\u80fd\u3001\u53ef\u7ef4\u62a4\u6027\u548c\u4ee3\u7801\u8d28\u91cf\u3002<\/li>\n<li><strong>\u589e\u52a0\u667a\u80fd\u4f53\u548c\u6a21\u578b\u8bbe\u7f6e<\/strong>\uff1a\u6211\u4eec\u8ba1\u5212\u8bc4\u4f30\u66f4\u591a\u5546\u4e1a\u667a\u80fd\u4f53\u3001\u667a\u80fd\u4f53-\u6a21\u578b\u914d\u7f6e\uff0c\u4ee5\u53ca\u5f00\u653e\u6743\u91cd\u6a21\u578b\uff0c\u4ee5\u4fbf\u56e2\u961f\u80fd\u591f\u6bd4\u8f83\u66f4\u5e7f\u6cdb\u7684\u8bbe\u7f6e\u3002<\/li>\n<\/ul>\n<p>\u8be5\u57fa\u51c6\u662f\u5f00\u653e\u7684\uff0c\u56e0\u6b64\u60a8\u53ef\u4ee5\u68c0\u67e5\u4efb\u52a1\u3001\u6bd4\u8f83\u7ed3\u679c\uff0c\u5e76\u544a\u8bc9\u6211\u4eec\u540e\u7eed\u5e94\u6db5\u76d6\u54ea\u4e9b Kotlin \u573a\u666f\u3002<\/p>\n\n\n<p><\/p>\n\n\n    <div class=\"about-author \">\n        <div class=\"about-author__box\">\n            <div class=\"row\">\n                <div class=\"about-author__box-img\">\n                    <img decoding=\"async\" src=\"https:\/\/blog.jetbrains.com\/wp-content\/uploads\/2021\/08\/photo_2021-08-03_15-27-43-200x200.jpg\" width=\"200\" height=\"200\" alt=\"Alyona Chernyaeva\" loading=\"lazy\"  class=\"avatar avatar-200 wp-user-avatar wp-user-avatar-200 photo avatar-default\">\n                <\/div>\n                <div class=\"about-author__box-text\">\n                                            <h4>Alyona Chernyaeva<\/h4>\n                                                        <\/div>\n            <\/div>\n        <\/div>\n    <\/div>\n","protected":false},"author":1297,"featured_media":725745,"comment_status":"closed","ping_status":"closed","template":"","categories":[8899],"tags":[],"cross-post-tag":[8396,6355],"class_list":["post-725730","kotlin","type-kotlin","status-publish","has-post-thumbnail","hentry","category-ai","cross-post-tag-ai","cross-post-tag-kotlin"],"acf":[],"_links":{"self":[{"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/kotlin\/725730","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/kotlin"}],"about":[{"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/types\/kotlin"}],"author":[{"embeddable":true,"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/users\/1297"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/comments?post=725730"}],"version-history":[{"count":2,"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/kotlin\/725730\/revisions"}],"predecessor-version":[{"id":725757,"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/kotlin\/725730\/revisions\/725757"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/media\/725745"}],"wp:attachment":[{"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/media?parent=725730"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/categories?post=725730"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/tags?post=725730"},{"taxonomy":"cross-post-tag","embeddable":true,"href":"https:\/\/blog.jetbrains.com\/zh-hans\/wp-json\/wp\/v2\/cross-post-tag?post=725730"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}