
{"id":8958,"date":"2026-06-09T19:11:55","date_gmt":"2026-06-09T11:11:55","guid":{"rendered":"https:\/\/infernews.com\/blog\/swe-explore-bench\/"},"modified":"2026-06-09T19:11:55","modified_gmt":"2026-06-09T11:11:55","slug":"swe-explore-bench","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/swe-explore-bench\/","title":{"rendered":"SWE-Explore-Bench\uff1a\u62c6\u89e3\u7de8\u78bc\u4ee3\u7406\u5982\u4f55\u7406\u89e3\u4f60\u7684\u7a0b\u5f0f\u78bc"},"content":{"rendered":"<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/06\/pasted-ec3b1f8ff966.jpg\" alt=\"SWE-Explore evaluates repository exploration directly instead of only end-to-end repair.\"><\/figure>\n<p>SWE-Explore-Bench \u7531\u4e0a\u6d77\u4ea4\u901a\u5927\u5b78\u3001\u9999\u6e2f\u4e2d\u6587\u5927\u5b78\u7b49\u5718\u968a\u63a8\u51fa\uff0c\u5c08\u9580\u8003\u6838\u7de8\u78bc\u4ee3\u7406\u5728\u300c\u771f\u6b63\u843d\u7b46\u4fee\u6539\u524d\u300d\u63a2\u7d22\u7a0b\u5f0f\u78bc\u5eab\u7684\u8868\u73fe\u3002\u73fe\u6709\u7684 SWE-bench \u7b49\u57fa\u6e96\u53ea\u7d66\u51fa\u6700\u7d42\u901a\u904e\u8207\u5426\u7684\u4e8c\u5143\u7d50\u679c\uff0c\u96e3\u4ee5\u5206\u8fa8\u4ee3\u7406\u662f\u56e0\u70ba\u627e\u5c0d\u4f4d\u7f6e\u800c\u6210\u529f\uff0c\u9084\u662f\u78b0\u5de7\u8499\u5c0d\u3002\u9019\u500b\u9805\u76ee\u628a\u63a2\u7d22\u9019\u4e00\u6b65\u62bd\u96e2\u51fa\u4f86\u55ae\u7368\u8a08\u5206\uff0c\u66f4\u8cbc\u8fd1\u8a3a\u65b7\u4ee3\u7406\u80fd\u529b\u7684\u672c\u8cea\u3002<\/p>\n<p>\u5177\u9ad4\u505a\u6cd5\u662f\u6536\u96c6\u540c\u4e00\u8b70\u984c\u7684\u591a\u689d\u6210\u529f\u4fee\u5fa9\u8ecc\u8de1\uff0c\u5f9e\u4e2d\u62bd\u53d6\u4ee3\u7406\u5be6\u969b\u8b80\u53d6\u7684\u7a0b\u5f0f\u78bc\u884c\u7bc4\u570d\uff0c\u6574\u5408\u51fa\u5171\u8b58\u7684\u6838\u5fc3\u4e0a\u4e0b\u6587\uff0c\u518d\u4fdd\u7559\u90e8\u5206\u6a21\u578b\u7368\u6709\u7684\u53ef\u9078\u4e0a\u4e0b\u6587\u3002\u4ee3\u7406\u9700\u8981\u8f38\u51fa\u4e00\u4efd\u6309\u884c\u7bc4\u570d\u6392\u5e8f\u7684\u7a0b\u5f0f\u78bc\u5340\u57df\u6e05\u55ae\uff0c\u8a55\u5206\u6db5\u84cb\u8986\u84cb\u7387\u3001\u6392\u5e8f\u54c1\u8cea\u3001\u4e0a\u4e0b\u6587\u6548\u7387\uff0c\u4ee5\u53ca\u4e0b\u6e38\u53d7\u9650\u4fee\u88dc\u9a57\u8b49\u56db\u500b\u7dad\u5ea6\u3002\u9019\u7a2e\u884c\u7d1a\u76e3\u7763\u6bd4\u50b3\u7d71\u7684\u6a94\u6848\u7d1a\u5b9a\u4f4d\u66f4\u7d30\u7dfb\uff0c\u80fd\u63ed\u793a\u4ee3\u7406\u7684\u771f\u6b63\u74f6\u9838\u3002<\/p>\n<p>\u8cc7\u6599\u96c6\u6db5\u84cb 10 \u7a2e\u7a0b\u5f0f\u8a9e\u8a00\u3001203 \u500b\u958b\u6e90\u9805\u76ee\u4e2d\u7684 848 \u500b\u8b70\u984c\uff0c\u4e26\u63d0\u4f9b OpenAI \u76f8\u5bb9\u7684\u7aef\u9ede\uff0c\u65b9\u4fbf\u63a5\u99c1\u4e0d\u540c LLM \u9032\u884c\u884c\u7bc4\u570d\u7cbe\u4fee\u3002\u5be6\u6e2c\u7d50\u679c\u986f\u793a\uff0c\u5177\u5099\u4ee3\u7406\u80fd\u529b\u7684\u63a2\u7d22\u5668\u660e\u986f\u9818\u5148\u50b3\u7d71\u6aa2\u7d22\u5668\uff0c\u73fe\u4ee3\u65b9\u6cd5\u5728\u6a94\u6848\u5c64\u7d1a\u5df2\u76f8\u7576\u6210\u719f\uff0c\u4f46\u884c\u7d1a\u8986\u84cb\u8207\u9ad8\u6548\u6392\u5e8f\u4ecd\u662f\u5340\u5206\u9802\u5c16\u65b9\u6848\u7684\u5206\u6c34\u5dba\u3002<\/p>\n<p>\u9069\u5408\u95dc\u6ce8 SWE-agent\u3001AutoCodeRover\u3001OpenHands \u7b49\u4ee3\u7406\u6846\u67b6\u7684\u7814\u7a76\u8005\u3001\u958b\u767c\u8005\u53ca\u57fa\u6e96\u8a2d\u8a08\u8005\u4f7f\u7528\u3002\u5c0d\u60f3\u4e86\u89e3\u81ea\u5bb6\u4ee3\u7406\u300c\u8b80\u78bc\u74b0\u7bc0\u300d\u5f37\u5f31\u7684\u5718\u968a\u800c\u8a00\uff0c\u9019\u662f\u4e00\u500b\u503c\u5f97\u653e\u5165\u8a55\u6e2c\u7ba1\u7dda\u7684\u53c3\u8003\u9805\u76ee\u3002<\/p>\n<p><strong>GitHub\uff1a<\/strong> <a href=\"https:\/\/github.com\/Qiushao-E\/SWE-Explore-Bench\" rel=\"noopener noreferrer\">https:\/\/github.com\/Qiushao-E\/SWE-Explore-Bench<\/a><\/p>\n<p><strong>Paper\uff1a<\/strong> <a href=\"https:\/\/arxiv.org\/pdf\/2606.07297\" rel=\"noopener noreferrer\">https:\/\/arxiv.org\/pdf\/2606.07297<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u9019\u500b\u57fa\u6e96\u5c07\u7de8\u78bc\u4ee3\u7406\u7684\u63a2\u7d22\u80fd\u529b\u7368\u7acb\u8a55\u5206\uff0c\u4ee5 848 \u500b\u771f\u5be6\u8b70\u984c\u70ba\u57fa\u790e\uff0c\u8861\u91cf\u4ee3\u7406\u80fd\u5426\u5728\u52d5\u624b\u4fee\u6539\u524d\u627e\u5c0d\u884c\u7d1a\u4e0a\u4e0b\u6587\u3002<\/p>\n","protected":false},"author":8,"featured_media":8957,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[133,164,146,197],"tags":[],"class_list":["post-8958","post","type-post","status-publish","format-standard","hentry","category-133","category-164","category-146","category-framework"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/8958","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/comments?post=8958"}],"version-history":[{"count":0,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/8958\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/8957"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=8958"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=8958"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=8958"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}