{"id":1446,"date":"2025-07-24T21:04:00","date_gmt":"2025-07-24T13:04:00","guid":{"rendered":"http:\/\/www.applicationsofllm.com\/?p=1446"},"modified":"2025-09-17T21:06:11","modified_gmt":"2025-09-17T13:06:11","slug":"%e3%80%90agi-eval%e8%af%84%e6%b5%8b%e6%8a%a5%e5%91%8a%e3%80%91%e6%b7%b1%e5%ba%a6%e6%8b%86%e8%a7%a3-qwen3%ef%bc%8c%e5%88%b7%e6%a6%9c%e4%b9%8b%e5%a4%96%ef%bc%8c%e6%9b%b4%e9%9c%80%e6%9e%84%e5%bb%ba","status":"publish","type":"post","link":"http:\/\/www.applicationsofllm.com\/index.php\/2025\/07\/24\/%e3%80%90agi-eval%e8%af%84%e6%b5%8b%e6%8a%a5%e5%91%8a%e3%80%91%e6%b7%b1%e5%ba%a6%e6%8b%86%e8%a7%a3-qwen3%ef%bc%8c%e5%88%b7%e6%a6%9c%e4%b9%8b%e5%a4%96%ef%bc%8c%e6%9b%b4%e9%9c%80%e6%9e%84%e5%bb%ba\/","title":{"rendered":"\u3010AGI-Eval\u8bc4\u6d4b\u62a5\u544a\u3011\u6df1\u5ea6\u62c6\u89e3 Qwen3\uff0c\u5237\u699c\u4e4b\u5916\uff0c\u66f4\u9700\u6784\u5efa\u5168\u9762\u8bc4\u6d4b\u65b0\u4f53\u7cfb"},"content":{"rendered":"\n<p>\u524d\u51e0\u5929\uff0c\u963f\u91cc\u5df4\u5df4\u6b63\u5f0f\u5f00\u6e90\u65b0\u4e00\u4ee3\u901a\u4e49\u5343\u95ee\u6a21\u578b Qwen3\u3002\u4e0a\u671f\u6211\u4eec\u4ece\u521b\u4f5c\u3001\u5206\u6790\u3001\u64cd\u4f5c\u548c\u95ee\u7b54\u56db\u65b9\u9762\u8fdb\u884c\u4e86\u4efb\u52a1\u5b9e\u6d4b\uff0c\u672c\u671f\u805a\u7126\u5176\u901a\u7528\u80fd\u529b\u4e0e\u5e94\u7528\u573a\u666f\u8bc4\u6d4b\u3002OpenAI \u5458\u5de5\u3001\u59da\u73ed\u6821\u53cb\u59da\u987a\u96e8\u53d1\u5e03\u6587\u7ae0\u300a\u5927\u6a21\u578b\u4e0b\u534a\u573a\u662f\u4ea7\u54c1\u7684\u6e38\u620f\uff0c\u505a\u7814\u7a76\u4e5f\u8981\u6709\u4ea7\u54c1\u601d\u7ef4\u300b\uff0c\u4eba\u5de5\u667a\u80fd\u7684\u91cd\u70b9\uff0c\u5df2\u7ecf\u4ece\u201c\u89e3\u51b3\u95ee\u9898\u201d\u8f6c\u5411\u4e86\u201c\u5b9a\u4e49\u95ee\u9898\u201d\uff0c\u8bc4\u4f30\u6216\u8bb8\u6bd4\u8bad\u7ec3\u66f4\u91cd\u8981\u3002\u6a21\u578b\u8bc4\u6d4b\u4e2d\uff0c\u5e94\u6452\u5f03 \u201c\u5237\u699c\u201d \u601d\u7ef4\uff0c\u4ee5\u5168\u9762\u8bc4\u6d4b\u89c6\u89d2\u7275\u5f15\u8bad\u7ec3\u53cd\u6620\u6a21\u578b\u771f\u5b9e\u6027\u80fd\uff0c\u8fdb\u800c\u63a8\u52a8\u6a21\u578b\u4f18\u5316\u3002<\/p>\n\n\n\n<p>\u76ee\u5f55\uff1a<\/p>\n\n\n\n<p>1.\u8bc4\u6d4b\u6838\u5fc3\u7ed3\u8bba<\/p>\n\n\n\n<p>2.\u901a\u7528\u8bc4\u6d4b\u7ed3\u679c<\/p>\n\n\n\n<p>3.\u5e94\u7528\u573a\u666f\u8bc4\u6d4b\u7ed3\u679c<\/p>\n\n\n\n<p>4.\u544a\u522b\u5237\u699c\uff0c\u6784\u5efa\u5168\u9762\u8bc4\u6d4b\u65b0\u4f53\u7cfb<\/p>\n\n\n\n<p><strong>1.\u8bc4\u6d4b\u6838\u5fc3\u7ed3\u8bba<\/strong><\/p>\n\n\n\n<p>Qwen3 \u6a21\u578b\u652f\u6301\u601d\u8003\u6a21\u5f0f\u548c\u975e\u601d\u8003\u6a21\u5f0f\uff0c\u53ef\u901a\u8fc7\u53c2\u6570\u5b9e\u73b0\u4e24\u79cd\u6a21\u5f0f\u7684\u5207\u6362\u3002Qwen3-235B-A22B\u7684\u5177\u4f53\u5b9e\u529b\u5982\u4f55\uff0cAGI-Eval \u5927\u6a21\u578b\u8bc4\u6d4b\u793e\u533a\u7b2c\u4e00\u65f6\u95f4\u505a\u4e86\u6a21\u578b\u8bc4\u6d4b\uff0c\u5ba2\u89c2\u8bc4\u6d4b\u7ed3\u8bba\u62a5\u544a\u5982\u4e0b\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>\u5728\u901a\u7528\u80fd\u529b\u4e0a<\/strong>\uff0cQwen3-235B-A22B\u00a0\u5728 Qwen3 \u6b64\u6b21\u66f4\u65b0\u6a21\u578b\u7cfb\u5217\u4e2d\u8868\u73b0\u6700\u5f3a\u3002\u5176\u4e2d<strong>Thinking \u6a21\u5f0f\u7565\u4f4e\u4e8e DeepSeek-R1\uff0c\u9ad8\u4e8e QwQ-32B\uff1b\u975e Thinking\u6a21\u5f0f\u843d\u540e\u4e8e DeepSeek-V3-0324\u3002<\/strong>\u540c\u65f6 Qwen3 \u7cfb\u5217\u5728 SimpleQA \u4e0a\u8868\u73b0\u76f8\u5bf9\u8f83\u5f31\uff0c<strong>\u5b58\u5728\u4e00\u5b9a\u5e7b\u89c9\u98ce\u9669\u3002<\/strong><\/li>\n\n\n\n<li><strong>\u5728\u8f93\u51fa\u957f\u5ea6\u4e0a<\/strong>\uff0c\u63a8\u7406\u4efb\u52a1\u4e2d\uff0cQwen3-235B-A22B \u7684\u5e73\u5747token\u4e3a3882\uff0c\u8f93\u51fa\u957f\u5ea6\u8f83\u00a0DeepSeek-R1 \uff08\u5e73\u5747token\u4e3a2977\uff09\u66f4\u957f\uff0c\u4f46\u76f8\u6bd4 QwQ-32B\uff08\u5e73\u5747token\u4e3a4596\uff09\u00a0\u6709\u7f29\u51cf\u3002<\/li>\n\n\n\n<li><strong>\u5728\u5e94\u7528\u573a\u666f\u4e0a<\/strong>\uff0cQwen3-32B\u3001Qwen3-235B-A22B\u6574\u4f53\u5904\u4e8e\u540c\u4e00\u6c34\u4f4d\uff0c\u6574\u4f53\u843d\u540eQwQ-32B\uff5e3.4-3.6pp\u3001\u843d\u540eDeepSeek-R1\uff5e10.3-10.4pp\u3002<\/li>\n<\/ul>\n\n\n\n<p>\u5173\u6ce8\u6211\u4eec\uff0c\u7b2c\u4e00\u65f6\u95f4\u83b7\u53d6\u6700\u65b0\u6a21\u578b\u8bc4\u6d4b\u5206\u6790<\/p>\n\n\n\n<p><strong>2.\u901a\u7528\u8bc4\u6d4b\u7ed3\u679c<\/strong><\/p>\n\n\n\n<p>\u8bc4\u6d4b<strong>\u57fa\u4e8e\u81ea\u5efa\u7684\u901a\u7528\u80fd\u529b\u8bc4\u6d4b\u96c6\u3001\u516c\u5f00\u8bc4\u6d4b\u96c6\u3002<\/strong><\/p>\n\n\n\n<p>\u81ea\u5efa\u8bc4\u6d4b\u96c6\u7531\u4e3b\u89c2\u3001\u5ba2\u89c2\u8bc4\u6d4b\u4e24\u90e8\u5206\u6784\u6210\uff0c\u8986\u76d6\u4ea4\u4e92\u80fd\u529b\u3001\u63a8\u7406\u3001\u77e5\u8bc6\u50a8\u5907\u7b49\u5404\u80fd\u529b\uff1b<br><br><\/p>\n\n\n\n<p>\u516c\u5f00\u8bc4\u6d4b\u96c6\u5747\u4e3a\u5ba2\u89c2\u8bc4\u6d4b\uff0c\u9009\u53d6\u4ee3\u7801\u3001\u590d\u6742\u63a8\u7406\u3001\u5b66\u79d1\u7b49\u5404\u7c7b\u6709\u5f71\u54cd\u529b\u548c\u533a\u5206\u5ea6\u7684\u8bc4\u6d4b\u96c6\uff0c\u5e76\u5bf9\u5176\u8fdb\u884c\u8d28\u68c0\u4fee\u6b63\uff0c\u91c7\u7528\u66f4\u5408\u7406\u7684\u8bc4\u6d4b\u65b9\u5f0f\uff08\u5982AIME\u7cfb\u5217\u8bc4\u6d4b\u96c6\u56e0\u672c\u8eab\u6ce2\u52a8\u8f83\u5927\uff0c\u91c7\u7528\u8dd110\u6b21\u7ed3\u679c\u53d6\u5747\u503c\u7684\u65b9\u5f0f\u8ba1\u7b97\u6307\u6807\uff0cIFEval_Pro\u4e3a\u793e\u533a\u57fa\u4e8eIFEval\u4fee\u6b63\u6539\u9020\u540e\u7684\u8bc4\u6d4b\u96c6\uff09\u3002<\/p>\n\n\n\n<p><strong>2.1 \u5185\u90e8\u8bc4\u6d4b\u7ed3\u679c<\/strong><\/p>\n\n\n\n<p>Qwen3-235B-A22B \u5728 Thinking \u6a21\u5f0f\u4e0b\u5ba2\u89c2\u51c6\u786e\u7387\u4e3a0.7469\uff0c\u7565\u4f4e\u4e8e DeepSeek-R1 \u76840.7660\uff0c\u9ad8\u4e8e QwQ-32B \u76840.6995\uff1b\u800c&nbsp;Qwen3-235B-A22B \u975eThinking \u6a21\u5f0f\u4e0e DeepSeek-V3-0324\u5dee\u8ddd\u8f83\u5927\uff08\u7ea64pp\uff09\u3002&nbsp;<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"445\" src=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-211-1024x445.png\" alt=\"\" class=\"wp-image-1451\" srcset=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-211-1024x445.png 1024w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-211-300x131.png 300w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-211-768x334.png 768w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-211.png 1108w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p>\u6b64\u5916\uff0c\u4ece<strong>\u8f93\u51fa\u957f\u5ea6<\/strong>\u4e0a\u6765\u770b\uff0c\u5728\u63a8\u7406\u4efb\u52a1\u4e0a\uff0cQwen3-235B-A22B \u7684\u5e73\u5747token\u4e3a3882\uff0c\u8f93\u51fa\u957f\u5ea6\u8f83 DeepSeek-R1\uff08\u5e73\u5747token\u4e3a2977\uff09\u66f4\u957f\uff0c\u4f46\u76f8\u6bd4 QwQ-32B\uff08\u5e73\u5747token\u4e3a4596\uff09 \u6709\u7f29\u51cf\u3002<\/p>\n\n\n\n<p><strong>2.2 \u516c\u5f00\u8bc4\u6d4b\u7ed3\u679c<\/strong><\/p>\n\n\n\n<p>Qwen3 \u7cfb\u5217\u5728 SimpleQA \u4e0a\u8868\u73b0\u76f8\u5bf9\u8f83\u5f31\uff0c<strong>\u5b58\u5728\u4e00\u5b9a\u7684\u5e7b\u89c9\u98ce\u9669<\/strong>\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"433\" src=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-209-1024x433.png\" alt=\"\" class=\"wp-image-1449\" srcset=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-209-1024x433.png 1024w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-209-300x127.png 300w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-209-768x324.png 768w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-209.png 1108w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p><strong>2.3 \u57fa\u5ea7\u6a21\u578b\u8bc4\u6d4b<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Qwen3 \u7cfb\u5217\u57fa\u5ea7\uff0c\u6700\u5927\u5c3a\u5bf8\u6a21\u578b\u672a\u5f00\u6e90\uff0c\u4ec5\u5f00\u6e90 Qwen3-30B-A3B-base\u3001Qwen3-14B-Base \u53ca\u4ee5\u4e0b\u5c3a\u5bf8\u3002<\/li>\n\n\n\n<li>\u5728\u901a\u7528\u57fa\u5ea7\u8bc4\u6d4b\u96c6\u4e0a\uff0cQwen3-30B-A3B-base \u548c Qwen3-14B-Base \u51c6\u786e\u7387\u4e0e\u8fd1\u671f\u5f00\u6e90\u7684Llama4-Maverick-17B-128E-\u57fa\u5ea7\u4ecd\u6709\u8f83\u5927\u5dee\u8ddd\u3002<\/li>\n<\/ul>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"593\" src=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-210-1024x593.png\" alt=\"\" class=\"wp-image-1450\" srcset=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-210-1024x593.png 1024w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-210-300x174.png 300w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-210-768x445.png 768w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-210.png 1108w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p><strong>3.\u5e94\u7528\u573a\u666f\u8bc4\u6d4b\u7ed3\u679c<\/strong><\/p>\n\n\n\n<p>\u6211\u4eec\u6765\u770b\u5728<strong>\u4ee3\u7801\u3001\u5546\u54c1\u548c\u670d\u52a1\u96f6\u552e\u3001\u751f\u6d3b\u767e\u79d1<\/strong>3\u4e2a\u5178\u578b\u5e94\u7528\u573a\u666f\u7684\u5ba2\u89c2\u8bc4\u6d4b\u96c6\u3002Qwen3-32B\u3001Qwen3-235B-A22B\uff08\u5f00\u542fThinking\u6a21\u5f0f\uff0c\u4e0b\u540c\uff09\u5927\u5e45\u843d\u540e\u4e8e\u6807\u6746\u63a8\u7406\u6a21\u578b DeepSeek-R1\uff0c\u751a\u81f3\u660e\u663e\u843d\u540e\u4e8e\u81ea\u5bb6\u4e0a\u4e00\u4ee3\u63a8\u7406\u6a21\u578b QwQ-32B\u3002<\/p>\n\n\n\n<p><strong>3.1 \u5e94\u7528\u573a\u666f\u6574\u4f53\u8bc4\u6d4b\u7ed3\u679c<\/strong><\/p>\n\n\n\n<p>Qwen3-32B \u548c Qwen3-235B-A22B \u5728\u5f00\u542f Thinking \u6a21\u5f0f\u4e0b\uff0c\u6574\u4f53\u5904\u4e8e\u540c\u4e00\u6c34\u4f4d\uff0c\u5bf9\u6bd4 QwQ-32B \u843d\u540e\uff5e3.4-3.6pp\uff0c\u5bf9\u6bd4 DeepSeek-R1 \u843d\u540e\uff5e10.3-10.4pp\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u5bf9\u6bd4 QwQ-32B\uff0cQwen3-32B \u548c Qwen3-235B-A22B \u6574\u4f53\u843d\u540e\uff5e3.4-3.6pp\uff0c\u5728\u4ee3\u7801\u573a\u666f\u4e0a\u5dee\u8ddd\u6700\u5927\uff08\u843d\u540e\uff5e7.7-9.3pp\uff09\uff0c\u5728\u5546\u54c1\u548c\u670d\u52a1\u96f6\u552e\u573a\u666f\u4e0a\u5dee\u8ddd\u4e3a4.2-5.0pp\uff0c\u5728\u751f\u6d3b\u767e\u79d1\u573a\u666f\u4e0a\u9886\u51481.1-4.0pp\uff1b<\/li>\n<\/ul>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"350\" src=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-207-1024x350.png\" alt=\"\" class=\"wp-image-1447\" srcset=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-207-1024x350.png 1024w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-207-300x102.png 300w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-207-768x262.png 768w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-207.png 1107w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p><em>\u25b3\u63a8\u7406\u6a21\u578b\u7684\u5bf9\u6bd4\u5f00\u542f\u4e86 Thinking \u6a21\u5f0f<\/em><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u5bf9\u6bd4 DeepSeek-R1\uff0cQwen3-32B \u548c Qwen3-235B-A22B \u6574\u4f53\u843d\u540e\uff5e10.3-10.4pp\uff0c\u5728\u4ee3\u7801\u3001\u5546\u54c1\u548c\u670d\u52a1\u96f6\u552e\u3001\u751f\u6d3b\u670d\u52a1\u573a\u666f\u4e0a\u5747\u843d\u540e\uff5e10pp\u3002<\/li>\n<\/ul>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"347\" src=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-208-1024x347.png\" alt=\"\" class=\"wp-image-1448\" srcset=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-208-1024x347.png 1024w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-208-300x102.png 300w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-208-768x260.png 768w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-208.png 1107w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p><strong>Qwen3\u7cfb\u5217\u5185\u90e8\u6765\u770b\uff1a<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Qwen3 \u7cfb\u5217\u7684\u76f8\u540c\u6a21\u578b\uff0c\u5f00\u542f Thinking \u6a21\u5f0f\u6bd4\u4e0d\u5f00\u542f Thinking \u6a21\u5f0f\u7684\u6307\u6807\u6574\u4f53\u9ad88.0-8.6pp\uff08\u5982\u4e0b\u8868\uff09\u3002<\/li>\n<\/ul>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"300\" src=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-213-1024x300.png\" alt=\"\" class=\"wp-image-1453\" srcset=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-213-1024x300.png 1024w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-213-300x88.png 300w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-213-768x225.png 768w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-213.png 1107w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p>\u5206\u573a\u666f\u6765\u770b\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Qwen3\u7cfb\u5217\u5728\u4ee3\u7801\u573a\u666f\u6574\u4f53\u8868\u73b0\u8f83\u5dee\uff0c\u843d\u540e\u4e8e QwQ-32B\u3001DeepSeek-R1 \u81f3\u5c117.7pp\u3002<\/strong>\u5176\u4e2d\uff0cQwen3-32B\u3001Qwen3-235B-A22B\u5728\u6307\u4ee4\u5408\u5e76\u573a\u666f\u4e0a\u8868\u73b0\u5c1a\u53ef\uff0c\u843d\u540eQwQ-32B 0.6-1.2pp\uff1b\u5728\u4ee3\u7801\u7ea0\u9519\u3001\u8865\u5168\u3001\u751f\u6210\u573a\u666f\u4e0a\u5927\u5e45\u843d\u540eQwQ-32B \u7ea67.0-20.2pp\uff1b<\/li>\n\n\n\n<li><strong>Qwen3\u7cfb\u5217\u5728\u5546\u54c1\u548c\u670d\u52a1\u96f6\u552e\u573a\u666f\u4e0a\u6574\u4f53\u8868\u73b0\u8f83\u5dee\uff0c\u843d\u540e\u4e8e QwQ-32B\u3001DeepSeek-R1\u81f3\u5c11 4.2pp\u3002<\/strong>\u5176\u4e2d\uff0cQwen3-32B\u3001Qwen3-235B-A22B\u5728\u9700\u8981\u903b\u8f91\u63a8\u7406\u7684\u8bc4\u6d4b\u96c6\u4e0a\u8868\u73b0\u8f83\u5dee\uff08\u5982\u5728\u641c\u7d22\u5e94\u7528\u573a\u666f\u76f8\u5173\u8bc4\u6d4b\u96c6\u4e0a\uff0c\u6613\u5c06\u4e0e\u7528\u6237\u641c\u7d22\u8bcd\/\u9700\u6c42\u4e0d\u76f8\u5173\u7684\u7ed3\u679c\u5224\u5b9a\u4e3a\u7b26\u5408\u9700\u6c42\uff0c\u6216\u9519\u5224\u641c\u7d22\u8bcd\u548c\u6539\u5199\u8bcd\u7684\u4e0a\/\u4e0b\u4f4d\u8bcd\u5173\u7cfb\uff09\uff1b\u5728\u5546\u54c1\u4fe1\u606f\u63d0\u53d6\u8fd9\u7c7b\u7b80\u5355\u4fe1\u606f\u5904\u7406\u4efb\u52a1\u4e0a\u8868\u73b0\u8f83\u597d\uff1b<\/li>\n\n\n\n<li><strong>Qwen3\u7cfb\u5217\u5728\u751f\u6d3b\u767e\u79d1\u573a\u666f\u6574\u4f53\u8868\u73b0\u4e0d\u9519\uff0c\u9886\u5148QwQ-32B\u7ea61.1-4.0pp\uff0c\u4f46\u8ddd\u79bbDeepSeek-R1\u8fd8\u6709\u81f3\u5c11 8.5pp\u5dee\u8ddd\u3002<\/strong>\u5176\u4e2d\uff0cQwen3-32B\u3001Qwen3-235B-A22B\u5728\u9700\u8981\u903b\u8f91\u63a8\u7406\u7684\u5730\u7406\u4f4d\u7f6e\u95ee\u7b54\u8bc4\u6d4b\u96c6\u4e0a\u8868\u73b0\u8f83\u5dee\uff0c\u5728\u8282\u65e5\u98df\u7269\u3001\u83dc\u7cfb\u7b49\u751f\u6d3b\u77e5\u8bc6\u8bc4\u6d4b\u96c6\u4e0a\u8868\u73b0\u8f83\u597d\u3002<\/li>\n<\/ul>\n\n\n\n<p>\u54b1\u4eec\u8be6\u7ec6\u770b\u770b Qwen3 \u7cfb\u5217\u91cd\u70b9\u5ba3\u4f20\u7684<strong>\u4ee3\u7801\u573a\u666f<\/strong>\u6307\u6807\u3002\u5bf9\u6bd4 Qwen3-32B\u3001DeepSeek-R1\uff0cQwen3 \u7cfb\u5217\u5dee\u8ddd\u660e\u663e\uff0c\u66f4\u6613\u51fa\u73b0\u8fb9\u754c\u60c5\u51b5\u8003\u8651\u4e0d\u5168\u7b49\u95ee\u9898\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"304\" src=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-214-1024x304.png\" alt=\"\" class=\"wp-image-1454\" srcset=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-214-1024x304.png 1024w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-214-300x89.png 300w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-214-768x228.png 768w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-214.png 1108w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"303\" src=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-215-1024x303.png\" alt=\"\" class=\"wp-image-1455\" srcset=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-215-1024x303.png 1024w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-215-300x89.png 300w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-215-768x227.png 768w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-215.png 1108w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p>\u8fd9\u91cc\u4e3a\u5927\u5bb6\u5c55\u793a Qwen3-32B \u5178\u578b\u9519\u8bef case\uff0c\u5728\u4e0b\u56fe\u6848\u4f8b\u4e2d\uff0cQwen3 \u6ca1\u6709\u8003\u8651\u5230\u5206\u6bcd\u4e3a0\u7684\u8fb9\u754c\u60c5\u51b5\uff0c\u5bfc\u81f4\u65e0\u6cd5\u901a\u8fc7\u5355\u5143\u6d4b\u8bd5\u3002<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"245\" src=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-216-1024x245.png\" alt=\"\" class=\"wp-image-1456\" srcset=\"http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-216-1024x245.png 1024w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-216-300x72.png 300w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-216-768x184.png 768w, http:\/\/www.applicationsofllm.com\/wp-content\/uploads\/2025\/09\/image-216.png 1107w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/figure>\n\n\n\n<p><em>\u25b3\u5de6\u4fa7\u4e3aQwQ-32B\uff0c\u53f3\u4fa7\u4e3aQwen3-32B<\/em><\/p>\n\n\n\n<p><strong>4.\u544a\u522b\u5237\u699c<\/strong><\/p>\n\n\n\n<p><strong>\u6784\u5efa\u5168\u9762\u8bc4\u6d4b\u65b0\u4f53\u7cfb<\/strong><\/p>\n\n\n\n<p>\u5728\u4eba\u5de5\u667a\u80fd\u6280\u672f\u9ad8\u901f\u8fed\u4ee3\u7684\u5f53\u4e0b\uff0cAI \u5f00\u53d1\u8005\u7fa4\u4f53\u6b63\u9762\u4e34\u4e00\u4e2a\u666e\u904d\u7684\u8ba4\u77e5\u8bef\u533a\uff1a\u5c06\u6a21\u578b\u5728\u7279\u5b9a\u6d4b\u8bd5\u4e2d\u7684\u8868\u73b0\u8bef\u8ba4\u4e3a\u771f\u6b63\u7684\u667a\u80fd\u3002\u8fd9\u4e00\u8ff7\u601d\u7684\u4ea7\u751f\uff0c\u4e0e\u884c\u4e1a\u5185\u76db\u884c\u7684 \u201c\u5237\u699c\u201d \u98ce\u6c14\u7d27\u5bc6\u76f8\u5173\u3002\u4ee5\u65b0\u4e00\u4ee3\u8bed\u8a00\u6a21\u578b Qwen3 \u4e3a\u4f8b\uff0c\u56f4\u7ed5 Qwen3 \u7684\u8ba8\u8bba\u5e76\u4e0d\u603b\u662f\u5168\u9762\u548c\u5ba2\u89c2\u7684\uff0c\u90e8\u5206\u8bc4\u4ef7\u505c\u7559\u5728\u8868\u9762\uff0c\u805a\u7126\u4e8e\u6a21\u578b\u5728\u699c\u5355\u4e0a\u7684\u6392\u540d\uff0c\u800c\u5ffd\u89c6\u4e86\u80cc\u540e\u7684\u6df1\u5c42\u903b\u8f91\u3002<br><br><\/p>\n\n\n\n<p><br><strong>\u5728\u6a21\u578b\u8bad\u7ec3\u7684\u8fc7\u7a0b\u4e2d\uff0c\u5168\u9762\u7684\u8bc4\u6d4b\u7275\u5f15\u81f3\u5173\u91cd\u8981<\/strong>\uff0c&nbsp;\u201c\u5237\u699c\u201d \u672c\u8d28\u4e0a\u662f\u4e00\u79cd\u901a\u8fc7\u9488\u5bf9\u6027\u8bad\u7ec3\u83b7\u53d6\u77ed\u671f\u6548\u5e94\u7684\u624b\u6bb5\u3002\u5f00\u53d1\u8005\u4e3a\u63d0\u5347\u6a21\u578b\u5728\u7279\u5b9a\u6d4b\u8bd5\u4e2d\u7684\u5206\u6570\uff0c\u6ca1\u6709\u4ece\u8db3\u591f\u7efc\u5408\u7684\u89d2\u5ea6\u6765\u8bc4\u4f30\u76f8\u5173\u56e0\u7d20\uff0c\u56f4\u7ed5\u6d4b\u8bd5\u4efb\u52a1\u7684\u89c4\u5219\u548c\u6570\u636e\u7279\u70b9\u8fdb\u884c\u4f18\u5316\uff0c\u4f7f\u6a21\u578b\u638c\u63e1\u5e94\u5bf9\u6d4b\u8bd5\u7684 \u201c\u4f2a\u80fd\u529b\u201d\u3002\u8fd9\u79cd \u201c\u4f2a\u80fd\u529b\u201d \u4e0e\u4eba\u7c7b\u667a\u80fd\u5728\u590d\u6742\u73b0\u5b9e\u573a\u666f\u4e2d\u5c55\u73b0\u51fa\u7684\u9002\u5e94\u6027\u3001\u521b\u9020\u529b\u5b58\u5728\u672c\u8d28\u5dee\u5f02\uff0c\u96be\u4ee5\u6a21\u62df\u771f\u5b9e\u4e16\u754c\u4e2d\u9519\u7efc\u590d\u6742\u7684\u5e94\u7528\u60c5\u5883\uff0c\u65e0\u6cd5\u771f\u5b9e\u53cd\u6620 Qwen3 \u5728\u5b9e\u9645\u5e94\u7528\u4e2d\u7684\u80fd\u529b\u4e0e\u6f5c\u529b\u3002<\/p>\n\n\n\n<p>Qwen3&nbsp;\u53ea\u5173\u6ce8\u4e86 AIME \u7b49\u9ad8\u96be\u5ea6\u516c\u5f00\u8bc4\u6d4b\uff0c\u5f15\u5165\u5e7b\u89c9\u98ce\u9669\uff0c<strong>\u5728\u5b9e\u9645\u5e94\u7528\u573a\u666f\u4e2d\uff0c\u5e76\u4e0d\u80fd\u6ee1\u8db3\u7528\u6237\u7684\u771f\u5b9e\u9700\u6c42<\/strong>\u3002\u5c06\u5176\u4e0e\u63a8\u7406\u7cfb\u6a21\u578b\u6807\u6746 DeepSeek &#8211; R1 \u5bf9\u6bd4\uff0c\u5176\u95f4\u5dee\u8ddd\u663e\u8457\uff1b\u590d\u4e0e\u8be5\u56e2\u961f\u4e0a\u4e00\u4ee3\u66f4\u5c0f\u5c3a\u5bf8\u7684\u6a21\u578b QwQ-32B \u76f8\u8f83\uff0c\u90e8\u5206\u771f\u5b9e\u5e94\u7528\u573a\u666f\u7adf\u5c1a\u4e0d\u53ca\u4e4b\u3002<\/p>\n\n\n\n<p>\u56e0\u6b64\uff0c\u5728\u4eba\u5de5\u667a\u80fd\u6a21\u578b\u7684\u8bc4\u6d4b\u4e2d\u9700\u6452\u5f03\u5237\u699c\u601d\u7ef4\u3002\u76f8\u8f83\u4e8e \u201c\u5237\u699c\u201d \u5e26\u6765\u7684\u77ed\u671f\u8868\u9762\u6210\u7ee9\uff0c\u6df1\u5165\u7684\u7275\u5f15\u8bad\u7ec3\u80fd\u591f\u5e2e\u52a9\u6211\u4eec\u53d1\u73b0\u6a21\u578b\u7684\u6f5c\u529b\uff0c\u8bc6\u522b\u5176\u5728\u957f\u8fdc\u53d1\u5c55\u4e2d\u7684\u4f18\u52bf\u4e0e\u4e0d\u8db3\u3002<strong>\u4e00\u4e2a\u5168\u9762\u7684\u8bc4\u6d4b\u4f53\u7cfb\u9700\u8981\u6db5\u76d6\u8bed\u4e49\u7406\u89e3\u3001\u751f\u6210\u8d28\u91cf\u3001\u7b97\u6cd5\u63a8\u7406\u3001\u5b9e\u9645\u4efb\u52a1\u6267\u884c\u591a\u4e2a\u7ef4\u5ea6\uff0c\u8fd8\u8981\u8003\u91cf\u6a21\u578b\u5728\u672a\u77e5\u9886\u57df\u7684\u8fc1\u79fb\u80fd\u529b\u3001\u5bf9\u6a21\u7cca\u4fe1\u606f\u7684\u5904\u7406\u667a\u6167\u3001\u4ee5\u53ca\u957f\u671f\u4ea4\u4e92\u4e2d\u7684\u7a33\u5b9a\u6027\u3002<\/strong>\u552f\u6709\u575a\u6301\u5168\u9762\u8bc4\u6d4b\u89c6\u89d2\uff0c\u4ee5\u771f\u5b9e\u573a\u666f\u7684\u7efc\u5408\u8868\u73b0\u4e3a\u5bfc\u5411\u8fdb\u884c\u7275\u5f15\u8bad\u7ec3\uff0c\u624d\u80fd\u53cd\u6620\u6a21\u578b\u7684\u771f\u5b9e\u6027\u80fd\uff0c\u63a8\u52a8\u5176\u6301\u7eed\u4f18\u5316\uff0c\u771f\u6b63\u6ee1\u8db3\u7528\u6237\u9700\u6c42 \u3002<\/p>\n\n\n\n<p><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u524d\u51e0\u5929\uff0c\u963f\u91cc\u5df4\u5df4\u6b63\u5f0f\u5f00\u6e90\u65b0\u4e00\u4ee3\u901a\u4e49\u5343\u95ee\u6a21\u578b Qwen3\u3002\u4e0a\u671f\u6211\u4eec\u4ece\u521b\u4f5c\u3001\u5206\u6790\u3001\u64cd\u4f5c\u548c\u95ee\u7b54\u56db\u65b9\u9762\u8fdb\u884c\u4e86\u4efb\u52a1\u5b9e\u6d4b\uff0c [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":1452,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[8],"tags":[],"class_list":["post-1446","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-8"],"_links":{"self":[{"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/posts\/1446","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/comments?post=1446"}],"version-history":[{"count":1,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/posts\/1446\/revisions"}],"predecessor-version":[{"id":1457,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/posts\/1446\/revisions\/1457"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/media\/1452"}],"wp:attachment":[{"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/media?parent=1446"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/categories?post=1446"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.applicationsofllm.com\/index.php\/wp-json\/wp\/v2\/tags?post=1446"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}