{"id":804084,"date":"2026-06-24T11:18:09","date_gmt":"2026-06-24T03:18:09","guid":{"rendered":"https:\/\/www.alphaseek.cn\/?p=804084"},"modified":"2026-06-24T11:18:09","modified_gmt":"2026-06-24T03:18:09","slug":"%e9%98%bf%e9%87%8c%e5%8d%83%e9%97%ae%e5%8f%91%e5%b8%83%e9%a6%96%e4%b8%aa%e5%8e%9f%e7%94%9f%e8%af%ad%e8%a8%80%e4%b8%96%e7%95%8c%e6%a8%a1%e5%9e%8bqwen-agentworld","status":"publish","type":"post","link":"https:\/\/www.alphaseek.cn\/?p=804084","title":{"rendered":"\u963f\u91cc\u5343\u95ee\u53d1\u5e03\u9996\u4e2a\u539f\u751f\u8bed\u8a00\u4e16\u754c\u6a21\u578bQwen-AgentWorld"},"content":{"rendered":"<p>IT\u4e4b\u5bb6 6 \u6708 24 \u65e5\u6d88\u606f\uff0c<a href=\"https:\/\/www.alphaseek.cn\/?tag=%e9%98%bf%e9%87%8c%e5%b7%b4%e5%b7%b4\" title=\"View all posts in \u963f\u91cc\u5df4\u5df4\" target=\"_blank\" class=\"tag_link\">\u963f\u91cc\u5df4\u5df4<\/a>\u5343\u95ee\u5927\u6a21\u578b\u4eca\u5929\u6b63\u5f0f\u53d1\u5e03 Qwen-AgentWorld\uff0c\u5b98\u65b9\u79f0\u8fd9\u662f\u9996\u4e2a\u539f\u751f\u8bed\u8a00\u4e16\u754c\u6a21\u578b\uff0c\u80fd\u591f\u5728\u4e03\u5927\u9886\u57df\u4e2d\u6a21\u62df\u667a\u80fd\u4f53\u4ea4\u4e92\u73af\u5883\uff0c\u63d0\u4f9b\u4e24\u79cd\u89c4\u6a21\uff0835B-A3B \u4e0e 397B-A17B\uff09\u3002<\/p>\n<p><strong>\u539f\u751f\u4e16\u754c\u5efa\u6a21<\/strong>\uff1a\u73af\u5883\u5efa\u6a21\u4ece\u7ee7\u7eed\u9884\u8bad\u7ec3\uff08CPT\uff09\u9636\u6bb5\u8d77\u5373\u4e3a\u8bad\u7ec3\u76ee\u6807\uff0c\u8d2f\u7a7f CPT \u2192 SFT \u2192 RL \u5168\u6d41\u7a0b\uff0c\u800c\u975e\u5bf9\u901a\u7528\u5927\u8bed\u8a00\u6a21\u578b\u7684\u4e8b\u540e\u9002\u914d\u3002<\/p>\n<p><strong>\u4e03\u5927\u9886\u57df\uff0c\u4e00\u4e2a\u6a21\u578b<\/strong>\uff1a\u5355\u4e00\u6a21\u578b\u540c\u65f6\u8986\u76d6\u6587\u672c\u7c7b\u73af\u5883\uff08MCP\u3001Search\u3001Terminal\u3001SWE\uff09\u4e0e GUI \u7c7b\u73af\u5883\uff08Web\u3001OS\u3001Android\uff09\uff0c\u5b9e\u73b0\u8de8\u9886\u57df\u77e5\u8bc6\u8fc1\u79fb\u3002<\/p>\n<p>\u540c\u6b65\u53d1\u5e03\u7684\u8fd8\u6709 AgentWorldBench\u2014\u2014 \u8986\u76d6\u4e03\u5927\u9886\u57df\u7684\u8bed\u8a00\u4e16\u754c\u6a21\u578b\u8bc4\u6d4b\u57fa\u51c6\uff0c\u6bcf\u6761\u6d4b\u8bd5\u6837\u672c\u5747\u914d\u5907\u771f\u5b9e\u73af\u5883\u6267\u884c\u6240\u5f97\u7684\u771f\u5b9e\u73af\u5883\u89c2\u6d4b\u6570\u636e\u3002\u6a21\u578b\u4e0e\u8bc4\u6d4b\u57fa\u51c6\u53ef\u4ece Hugging Face \u548c ModelScope \u83b7\u53d6\u3002<\/p>\n<p>IT\u4e4b\u5bb6\u9644\u5b98\u65b9\u4ecb\u7ecd\u5982\u4e0b\uff1a<\/p>\n<p><strong>Qwen-AgentWorld<\/strong>\u4ee3\u8868\u4e86\u6211\u4eec\u7684\u6838\u5fc3\u63a2\u7d22\uff1a\u57fa\u4e8e\u8bed\u8a00\u6a21\u578b\u7684\u4e16\u754c\u5efa\u6a21\uff0c\u80fd\u5426\u8fdb\u4e00\u6b65\u62d3\u5c55\u901a\u7528\u667a\u80fd\u4f53\u80fd\u529b\u7684\u8fb9\u754c\u3002<\/p>\n<p>\u6211\u4eec\u4ece\u4e24\u4e2a\u65b9\u5411\u63a2\u7d22\u5982\u4f55\u5b9e\u73b0\u8bed\u8a00\u4e16\u754c\u5efa\u6a21\uff0c\u4ee5\u53ca\u5982\u4f55\u5c06\u5176\u5e94\u7528\u4e8e\u63a8\u8fdb\u901a\u7528\u667a\u80fd\u4f53\uff1a<\/p>\n<p>\u9996\u5148\uff0c\u6211\u4eec<strong>\u6784\u5efa\u4e86\u667a\u80fd\u4f53\u73af\u5883\u6a21\u62df\u7684\u57fa\u7840\u6a21\u578b<\/strong>\uff1aQwen-AgentWorld \u662f\u9996\u4e2a\u5728\u5355\u4e00\u6a21\u578b\u4e2d\u8986\u76d6\u4e03\u5927\u667a\u80fd\u4f53\u4ea4\u4e92\u9886\u57df\uff08MCP\u3001Search\u3001Terminal\u3001SWE\u3001Web\u3001OS\u3001Android\uff09\u7684\u8bed\u8a00\u4e16\u754c\u6a21\u578b\uff0c\u57fa\u4e8e\u8d85\u8fc7 1000 \u4e07\u6761\u771f\u5b9e\u73af\u5883\u4ea4\u4e92\u8f68\u8ff9\uff0c\u7ecf\u7531 CPT \u2192 SFT \u2192 RL \u4e09\u9636\u6bb5\u8bad\u7ec3\u800c\u6210\u3002\u5728 AgentWorldBench \u8bc4\u6d4b\u4e2d\uff0cQwen-AgentWorld-397B-A17B \u53d6\u5f97\u4e86\u6700\u9ad8\u7684\u6574\u4f53\u6a21\u62df\u8d28\u91cf\uff0c\u8d85\u8d8a GPT-5.4\u3001Claude Opus 4.8 \u4e0e Gemini 3.1 Pro\u3002<\/p>\n<p>\u5176\u6b21\uff0c\u6211\u4eec<strong>\u63a2\u8ba8\u4e16\u754c\u5efa\u6a21\u5728\u667a\u80fd\u4f53\u8bad\u7ec3\u4e2d\u7684\u4f5c\u7528<\/strong>\uff0c\u5e76\u901a\u8fc7\u4e24\u79cd\u4e92\u8865\u8303\u5f0f\u52a0\u4ee5\u9a8c\u8bc1\uff1a\u4f5c\u4e3a\u89e3\u8026\u7684\u73af\u5883\u6a21\u62df\u5668\uff0c\u5b83\u4e3a\u667a\u80fd\u4f53\u5f3a\u5316\u5b66\u4e60\u63d0\u4f9b\u4e86\u66f4\u4f18\u7684\u53ef\u6269\u5c55\u6027\u4e0e\u53ef\u63a7\u6027 \u2014\u2014 \u53ef\u63a7\u7684\u6a21\u62df RL \u80fd\u591f\u4ee5\u771f\u5b9e\u73af\u5883\u65e0\u6cd5\u5b9e\u73b0\u7684\u65b9\u5f0f\u5851\u9020\u667a\u80fd\u4f53\u884c\u4e3a\uff0c\u4e14\u663e\u8457\u4f18\u4e8e\u4ec5\u5728\u771f\u5b9e\u73af\u5883\u4e2d\u8bad\u7ec3\u7684 RL\uff1b\u4f5c\u4e3a\u7edf\u4e00\u7684\u667a\u80fd\u4f53\u57fa\u7840\u6a21\u578b\uff0cLWM \u9884\u70ed\u8bad\u7ec3\u53ef\u6709\u6548\u8fc1\u79fb\u81f3\u6db5\u76d6\u4e03\u4e2a\u57fa\u51c6\uff08\u5176\u4e2d\u4e09\u4e2a\u5b8c\u5168\u672a\u51fa\u73b0\u5728\u8bad\u7ec3\u96c6\u4e2d\uff09\u7684\u591a\u8f6e\u667a\u80fd\u4f53\u4efb\u52a1\uff0c\u4e14\u65e0\u9700\u5728\u667a\u80fd\u4f53\u4efb\u52a1\u4e0a\u8fdb\u884c\u4efb\u4f55 RL \u5fae\u8c03\uff0c\u521d\u6b65\u9a8c\u8bc1\u4e86\u8bed\u8a00\u4e16\u754c\u6a21\u578b\u80fd\u591f\u4f5c\u4e3a\u6784\u5efa\u66f4\u5f3a\u667a\u80fd\u4f53\u6a21\u578b\u7684\u57fa\u7840\u3002<\/p>\n<p>\u901a\u8fc7\u4e09\u9636\u6bb5\u8bad\u7ec3\u8303\u5f0f \u2014\u2014 \u6301\u7eed\u9884\u8bad\u7ec3\uff08CPT\uff09\u6ce8\u5165\u73af\u5883\u77e5\u8bc6\u3001\u76d1\u7763\u5fae\u8c03\uff08SFT\uff09\u6fc0\u6d3b\u4e0b\u4e00\u72b6\u6001\u9884\u6d4b\u63a8\u7406\u3001\u5f3a\u5316\u5b66\u4e60\uff08RL\uff09\u6253\u78e8\u6a21\u62df\u771f\u5b9e\u6027 \u2014\u2014 \u81ea\u5e95\u5411\u4e0a\u9010\u6b65\u6784\u5efa\u4e16\u754c\u5efa\u6a21\u80fd\u529b\u3002\u6211\u4eec\u63a2\u7d22\u4e86\u4e16\u754c\u6a21\u578b\u8d4b\u80fd\u901a\u7528\u667a\u80fd\u4f53\u7684\u4e24\u79cd\u4e92\u8865\u8303\u5f0f\uff1a\u4f5c\u4e3a\u89e3\u8026\u7684\u73af\u5883\u6a21\u62df\u5668\uff0c\u6211\u4eec\u5728 Tool Decathlon\u3001MCPMark \u548c WideSearch \u4e0a\u9a8c\u8bc1\u4e86\u53ef\u63a7\u6a21\u62df\u7684\u6709\u6548\u6027\uff0c\u5176\u8868\u73b0\u8d85\u8d8a\u4e86\u65e0\u63a7\u6a21\u62df\u4e0e\u771f\u5b9e\u73af\u5883\u8bad\u7ec3\uff1b\u4f5c\u4e3a\u7edf\u4e00\u667a\u80fd\u4f53\u57fa\u7840\u6a21\u578b\uff0c\u8bed\u8a00\u4e16\u754c\u6a21\u578b\uff08LWM\uff09\u7684\u9884\u70ed\u8bad\u7ec3\u53ef\u8fc1\u79fb\u81f3\u6db5\u76d6\u4e03\u4e2a\u57fa\u51c6\uff08\u5176\u4e2d\u4e09\u4e2a\u5b8c\u5168\u5c5e\u4e8e\u57df\u5916\uff09\u7684\u591a\u8f6e\u667a\u80fd\u4f53\u4efb\u52a1\uff0c\u521d\u6b65\u9a8c\u8bc1\u4e86\u8bed\u8a00\u4e16\u754c\u6a21\u578b\u80fd\u591f\u4f5c\u4e3a\u6784\u5efa\u66f4\u5f3a\u667a\u80fd\u4f53\u6a21\u578b\u7684\u57fa\u7840\u3002\u8bed\u8a00\u4e16\u754c\u5efa\u6a21\u5f00\u8f9f\u4e86\u4e00\u6761\u4e92\u8865\u7684\u6269\u5c55\u8def\u5f84\uff0c\u4f7f\u901a\u7528\u667a\u80fd\u4f53\u7684\u80fd\u529b\u8fb9\u754c\u5f97\u4ee5\u7a81\u7834\u771f\u5b9e\u73af\u5883\u4ea4\u4e92\u6240\u80fd\u63d0\u4f9b\u7684\u4e0a\u9650\u3002<\/p>\n<p>AgentWorldBench \u5df2\u5728 Hugging Face \u4e0a\u53d1\u5e03\uff0c\u4ee5\u6309\u9886\u57df\u5212\u5206\u7684 JSONL \u6587\u4ef6\u5f62\u5f0f\u63d0\u4f9b\uff0c\u6bcf\u4e2a\u6587\u4ef6\u5305\u542b\u6765\u81ea\u771f\u5b9e\u73af\u5883\u7684\u4ea4\u4e92\u8f68\u8ff9\u53ca\u5bf9\u5e94\u7684\u771f\u5b9e\u89c2\u6d4b\u6570\u636e\u3002<\/p>\n<p>(\u6765\u6e90\uff1a\u65b0\u6d6a\u79d1\u6280)<\/p>\n","protected":false},"excerpt":{"rendered":"<p>IT\u4e4b\u5bb6 6 \u6708 24 \u65e5\u6d88\u606f\uff0c\u963f\u91cc\u5df4\u5df4\u5343\u95ee\u5927\u6a21\u578b\u4eca\u5929\u6b63\u5f0f\u53d1\u5e03 Qwen-AgentWorld\uff0c\u5b98\u65b9\u79f0\u8fd9\u662f\u9996\u4e2a [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"aside","meta":[],"categories":[6],"tags":[121],"views":3,"_links":{"self":[{"href":"https:\/\/www.alphaseek.cn\/index.php?rest_route=\/wp\/v2\/posts\/804084"}],"collection":[{"href":"https:\/\/www.alphaseek.cn\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.alphaseek.cn\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.alphaseek.cn\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.alphaseek.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=804084"}],"version-history":[{"count":0,"href":"https:\/\/www.alphaseek.cn\/index.php?rest_route=\/wp\/v2\/posts\/804084\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.alphaseek.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=804084"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.alphaseek.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=804084"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.alphaseek.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=804084"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}