{"id":53788,"date":"2025-02-16T08:04:17","date_gmt":"2025-02-16T00:04:17","guid":{"rendered":"https:\/\/fwq.ai\/blog\/53788\/"},"modified":"2025-02-16T08:04:17","modified_gmt":"2025-02-16T00:04:17","slug":"deepseek-r1-vs-v3%ef%bc%9a%e5%a6%82%e4%bd%95%e9%80%89%e6%8b%a9%ef%bc%9f","status":"publish","type":"post","link":"https:\/\/fwq.ai\/blog\/53788\/","title":{"rendered":"DeepSeek R1 vs. V3\uff1a\u5982\u4f55\u9009\u62e9\uff1f"},"content":{"rendered":"

\u5728\u624b\u673a\u6216\u684c\u9762\u4e0a\u4f7f\u7528 DeepSeek \u5e94\u7528\u7a0b\u5e8f\u65f6\uff0c\u6211\u4eec\u53ef\u80fd\u4f1a\u4e0d\u786e\u5b9a\u4f55\u65f6\u9009\u62e9 R1\uff08\u4e5f\u79f0\u4e3a DeepThink\uff09\uff0c\u800c\u4e0d\u662f\u65e5\u5e38\u4efb\u52a1\u7684\u9ed8\u8ba4 V3 \u6a21\u578b\u3002<\/p>\n

\u5bf9\u4e8e\u5f00\u53d1\u4eba\u5458\u6765\u8bf4\uff0c\u6311\u6218\u6709\u70b9\u4e0d\u540c\u3002\u5f53\u901a\u8fc7\u5176 API \u96c6\u6210 DeepSeek \u65f6\uff0c\u6311\u6218\u5728\u4e8e\u627e\u51fa\u54ea\u79cd\u6a21\u578b\u66f4\u7b26\u5408\u6211\u4eec\u7684\u9879\u76ee\u8981\u6c42\u5e76\u589e\u5f3a\u529f\u80fd\u3002<\/p>\n

\u5728\u6b64\u535a\u5ba2\u4e2d\uff0c\u6211\u5c06\u4ecb\u7ecd\u8fd9\u4e24\u79cd\u6a21\u578b\u7684\u5173\u952e\u65b9\u9762\uff0c\u4ee5\u5e2e\u52a9\u4f60\u66f4\u8f7b\u677e\u5730\u505a\u51fa\u8fd9\u4e9b\u9009\u62e9\u3002\u6211\u5c06\u63d0\u4f9b\u793a\u4f8b\u6765\u8bf4\u660e\u6bcf\u4e2a\u6a21\u578b\u5728\u4e0d\u540c\u60c5\u51b5\u4e0b\u7684\u884c\u4e3a\u548c\u6027\u80fd\u3002\u6211\u8fd8\u4f1a\u4e3a\u4f60\u63d0\u4f9b\u4e00\u4e2a\u51b3\u7b56\u6307\u5357\uff0c\u53ef\u4ee5\u4f7f\u7528\u5b83\u6765\u5728 DeepSeek-R1 \u548c DeepSeek-V3 \u4e4b\u95f4\u8fdb\u884c\u9009\u62e9\u3002<\/p>\n

1\u3001DeepSeek-V3 \u548c DeepSeek-R1<\/h2>\n

DeepSeek \u662f\u4e00\u5bb6\u4e2d\u56fd AI \u521d\u521b\u516c\u53f8\uff0c\u5b83\u4ee5\u6bd4 OpenAI \u7684 o1 \u4f4e\u5f97\u591a\u7684\u6210\u672c\u5f00\u53d1\u4e86 DeepSeek-R1\uff0c\u5f15\u8d77\u4e86\u56fd\u9645\u5173\u6ce8\u3002\u5c31\u50cf OpenAI \u6709\u6211\u4eec\u90fd\u77e5\u9053\u7684 ChatGPT \u5e94\u7528\u7a0b\u5e8f\u4e00\u6837\uff0cDeepSeek \u4e5f\u6709\u7c7b\u4f3c\u7684\u804a\u5929\u673a\u5668\u4eba\uff0c\u5b83\u6709\u4e24\u4e2a\u6a21\u578b\uff1aDeepSeek-V3 \u548c DeepSeek-R1\u3002<\/p>\n

1.1 \u4ec0\u4e48\u662f DeepSeek-V3\uff1f<\/h3>\n

DeepSeek-V3 \u662f\u6211\u4eec\u4e0e DeepSeek \u5e94\u7528\u7a0b\u5e8f\u4ea4\u4e92\u65f6\u4f7f\u7528\u7684\u9ed8\u8ba4\u6a21\u578b\u3002\u5b83\u662f\u4e00\u79cd\u591a\u529f\u80fd\u7684\u5927\u578b\u8bed\u8a00\u6a21\u578b (LLM)\uff0c\u662f\u4e00\u79cd\u53ef\u4ee5\u5904\u7406\u5404\u79cd\u4efb\u52a1\u7684\u901a\u7528\u5de5\u5177\u3002<\/p>\n

\u8be5\u6a21\u578b\u4e0e\u5176\u4ed6\u77e5\u540d\u8bed\u8a00\u6a21\u578b\uff08\u5982 OpenAI \u7684 GPT-4o\uff09\u7ade\u4e89\u3002<\/p>\n

DeepSeek-V3 \u7684\u4e3b\u8981\u529f\u80fd\u4e4b\u4e00\u662f\u5b83\u4f7f\u7528\u4e86\u6df7\u5408\u4e13\u5bb6 (MoE) \u65b9\u6cd5\u3002\u8fd9\u79cd\u65b9\u6cd5\u5141\u8bb8\u6a21\u578b\u4ece\u4e0d\u540c\u7684\u201c\u4e13\u5bb6\u201d\u4e2d\u8fdb\u884c\u9009\u62e9\u6765\u6267\u884c\u7279\u5b9a\u4efb\u52a1\u3002\u5728\u4f60\u5411\u6a21\u578b\u63d0\u4f9b\u63d0\u793a\u540e\uff0c\u53ea\u6709\u6a21\u578b\u4e2d\u6700\u76f8\u5173\u7684\u90e8\u5206\u624d\u4f1a\u9488\u5bf9\u4efb\u4f55\u7ed9\u5b9a\u4efb\u52a1\u5904\u4e8e\u6d3b\u52a8\u72b6\u6001\uff0c\u4ece\u800c\u8282\u7701\u8ba1\u7b97\u8d44\u6e90\u5e76\u63d0\u4f9b\u7cbe\u786e\u7684\u7ed3\u679c\u3002<\/p>\n

\u672c\u8d28\u4e0a\uff0cDeepSeek-V3 \u662f\u6211\u4eec\u9700\u8981 LLM \u5b8c\u6210\u7684\u5927\u591a\u6570\u65e5\u5e38\u4efb\u52a1\u7684\u53ef\u9760\u9009\u62e9\u3002\u4f46\u662f\uff0c\u4e0e\u5927\u591a\u6570 LLM \u4e00\u6837\uff0c\u5b83\u4f7f\u7528\u4e0b\u4e00\u4e2a\u5355\u8bcd\u9884\u6d4b\u6765\u5de5\u4f5c\uff0c\u8fd9\u9650\u5236\u4e86\u5b83\u89e3\u51b3\u9700\u8981\u63a8\u7406\u7684\u95ee\u9898\u6216\u63d0\u51fa\u8bad\u7ec3\u6570\u636e\u4e2d\u672a\u4ee5\u67d0\u79cd\u65b9\u5f0f\u7f16\u7801\u7684\u65b0\u7b54\u6848\u7684\u80fd\u529b\u3002<\/p>\n

1.2 \u4ec0\u4e48\u662f DeepSeek-R1\uff1f<\/h3>\n

DeepSeek-R1 \u662f\u4e00\u4e2a\u5f3a\u5927\u7684\u63a8\u7406\u6a21\u578b\uff0c\u4e13\u4e3a\u89e3\u51b3\u9700\u8981\u9ad8\u7ea7\u63a8\u7406\u548c\u6df1\u5ea6\u89e3\u51b3\u95ee\u9898\u7684\u4efb\u52a1\u800c\u6784\u5efa\u3002\u5b83\u975e\u5e38\u9002\u5408\u89e3\u51b3\u7f16\u7801\u6311\u6218\uff0c\u8fd9\u4e9b\u6311\u6218\u4e0d\u4ec5\u4ec5\u662f\u91cd\u590d\u7f16\u5199\u4e86\u6570\u5343\u6b21\u7684\u4ee3\u7801\u548c\u903b\u8f91\u7e41\u91cd\u7684\u95ee\u9898\u3002<\/p>\n

\u5f53\u8981\u89e3\u51b3\u7684\u4efb\u52a1\u9700\u8981\u9ad8\u7ea7\u8ba4\u77e5\u64cd\u4f5c\uff08\u7c7b\u4f3c\u4e8e\u4e13\u4e1a\u6216\u4e13\u5bb6\u7ea7\u63a8\u7406\uff09\u65f6\uff0c\u53ef\u4ee5\u5c06\u5176\u89c6\u4e3a\u4f60\u7684\u9996\u9009\u3002<\/p>\n

\u6211\u4eec\u901a\u8fc7\u5355\u51fb\u201cDeepThink (R1)\u201d\u6309\u94ae\u6765\u6fc0\u6d3b\u5b83\uff1a<\/p>\n

DeepSeek-R1 \u7684\u4e0e\u4f17\u4e0d\u540c\u4e4b\u5904\u5728\u4e8e\u5b83\u5bf9\u5f3a\u5316\u5b66\u4e60\u7684\u7279\u6b8a\u4f7f\u7528\u3002\u4e3a\u4e86\u8bad\u7ec3 R1\uff0cDeepSeek \u5728 V3 \u5960\u5b9a\u7684\u57fa\u7840\u4e0a\u6784\u5efa\uff0c\u5229\u7528\u5176\u5e7f\u6cdb\u7684\u529f\u80fd\u548c\u5de8\u5927\u7684\u53c2\u6570\u7a7a\u95f4\u3002\u4ed6\u4eec\u901a\u8fc7\u5141\u8bb8\u6a21\u578b\u4e3a\u89e3\u51b3\u95ee\u9898\u7684\u573a\u666f\u751f\u6210\u5404\u79cd\u89e3\u51b3\u65b9\u6848\u6765\u6267\u884c\u5f3a\u5316\u5b66\u4e60\u3002\u7136\u540e\u4f7f\u7528\u57fa\u4e8e\u89c4\u5219\u7684\u5956\u52b1\u7cfb\u7edf\u6765\u8bc4\u4f30\u7b54\u6848\u548c\u63a8\u7406\u6b65\u9aa4\u7684\u6b63\u786e\u6027\u3002\u8fd9\u79cd\u5f3a\u5316\u5b66\u4e60\u65b9\u6cd5\u9f13\u52b1\u6a21\u578b\u968f\u7740\u65f6\u95f4\u7684\u63a8\u79fb\u5b8c\u5584\u5176\u63a8\u7406\u80fd\u529b\uff0c\u6709\u6548\u5730\u5b66\u4e60\u81ea\u4e3b\u63a2\u7d22\u548c\u5f00\u53d1\u63a8\u7406\u8def\u5f84\u3002<\/p>\n

DeepSeek-R1 \u662f OpenAI \u7684 o1 \u7684\u76f4\u63a5\u7ade\u4e89\u5bf9\u624b\u3002<\/p>\n

V3 \u548c R1 \u4e4b\u95f4\u7684\u4e00\u4e2a\u533a\u522b\u662f\uff0c\u5f53\u4e0e R1 \u804a\u5929\u65f6\uff0c\u6211\u4eec\u4e0d\u4f1a\u7acb\u5373\u5f97\u5230\u56de\u590d\u3002\u8be5\u6a21\u578b\u9996\u5148\u4f7f\u7528\u601d\u7ef4\u94fe\u63a8\u7406\u6765\u601d\u8003\u95ee\u9898\u3002\u53ea\u6709\u5f53\u5b83\u5b8c\u6210\u601d\u8003\u540e\uff0c\u5b83\u624d\u4f1a\u5f00\u59cb\u8f93\u51fa\u7b54\u6848\u3002<\/p>\n

\u8fd9\u4e5f\u610f\u5473\u7740\uff0c\u4e00\u822c\u6765\u8bf4\uff0cR1 \u7684\u54cd\u5e94\u901f\u5ea6\u6bd4 V3 \u6162\u5f97\u591a\uff0c\u56e0\u4e3a\u601d\u8003\u8fc7\u7a0b\u53ef\u80fd\u9700\u8981\u51e0\u5206\u949f\u624d\u80fd\u5b8c\u6210\uff0c\u6211\u4eec\u5c06\u5728\u540e\u9762\u7684\u4f8b\u5b50\u4e2d\u770b\u5230\u3002<\/p>\n

2\u3001V3 \u548c R1 \u4e4b\u95f4\u7684\u5dee\u5f02<\/h2>\n

\u8ba9\u6211\u4eec\u4ece\u5404\u4e2a\u65b9\u9762\u6765\u770b\u4e00\u4e0b DeepSeek-R1 \u548c DeepSeek-V3 \u4e4b\u95f4\u7684\u5dee\u5f02\uff1a<\/p>\n

2.1 \u63a8\u7406\u80fd\u529b<\/h3>\n

DeepSeek-V3 \u6ca1\u6709\u63a8\u7406\u80fd\u529b\u3002\u6b63\u5982\u6211\u4eec\u6240\u63d0\u5230\u7684\uff0c\u5b83\u5145\u5f53\u4e0b\u4e00\u4e2a\u5355\u8bcd\u9884\u6d4b\u5668\u3002\u8fd9\u610f\u5473\u7740\u5b83\u53ef\u4ee5\u56de\u7b54\u7b54\u6848\u7f16\u7801\u5728\u8bad\u7ec3\u6570\u636e\u4e2d\u7684\u95ee\u9898\u3002<\/p>\n

\u7531\u4e8e\u7528\u4e8e\u8bad\u7ec3\u8fd9\u4e9b\u6a21\u578b\u7684\u6570\u636e\u91cf\u975e\u5e38\u5e9e\u5927\uff0c\u56e0\u6b64\u5b83\u51e0\u4e4e\u53ef\u4ee5\u56de\u7b54\u4efb\u4f55\u4e3b\u9898\u7684\u95ee\u9898\u3002\u4e0e\u5176\u4ed6 LLM \u4e00\u6837\uff0c\u5b83\u5728\u81ea\u7136\u5bf9\u8bdd\u548c\u521b\u9020\u529b\u65b9\u9762\u8868\u73b0\u51fa\u8272\u3002\u8fd9\u662f\u6211\u4eec\u60f3\u8981\u7684\u6a21\u578b\uff0c\u7528\u4e8e\u521b\u4f5c\u6587\u7ae0\u3001\u5185\u5bb9\u521b\u4f5c\u6216\u56de\u7b54\u53ef\u80fd\u5df2\u7ecf\u88ab\u89e3\u51b3\u8fc7\u65e0\u6570\u6b21\u7684\u4e00\u822c\u95ee\u9898\u3002<\/p>\n

\u53e6\u4e00\u65b9\u9762\uff0cDeepSeek-R1 \u5728\u89e3\u51b3\u590d\u6742\u7684\u95ee\u9898\u3001\u903b\u8f91\u548c\u9010\u6b65\u63a8\u7406\u4efb\u52a1\u65f6\u8868\u73b0\u51fa\u8272\u3002\u5b83\u65e8\u5728\u89e3\u51b3\u9700\u8981\u5f7b\u5e95\u5206\u6790\u548c\u7ed3\u6784\u5316\u89e3\u51b3\u65b9\u6848\u7684\u5177\u6709\u6311\u6218\u6027\u7684\u67e5\u8be2\u3002\u5f53\u9762\u5bf9\u590d\u6742\u7684\u7f16\u7801\u6311\u6218\u6216\u8be6\u7ec6\u7684\u903b\u8f91\u8c1c\u9898\u65f6\uff0cR1 \u662f\u503c\u5f97\u4fe1\u8d56\u7684\u5de5\u5177\u3002<\/p>\n

2.2 \u901f\u5ea6\u548c\u6548\u7387<\/h3>\n

DeepSeek-V3 \u53d7\u76ca\u4e8e\u5176\u6df7\u5408\u4e13\u5bb6 (MoE) \u67b6\u6784\uff0c\u4f7f\u5176\u80fd\u591f\u66f4\u5feb\u3001\u66f4\u6709\u6548\u5730\u505a\u51fa\u54cd\u5e94\u3002\u8fd9\u4f7f\u5f97 V3 \u9002\u7528\u4e8e\u901f\u5ea6\u81f3\u5173\u91cd\u8981\u7684\u5b9e\u65f6\u4ea4\u4e92\u3002<\/p>\n

DeepSeek-R1 \u901a\u5e38\u9700\u8981\u66f4\u957f\u7684\u65f6\u95f4\u6765\u751f\u6210\u54cd\u5e94\uff0c\u4f46\u8fd9\u662f\u56e0\u4e3a\u5b83\u4e13\u6ce8\u4e8e\u63d0\u4f9b\u66f4\u6df1\u5165\u3001\u66f4\u7ed3\u6784\u5316\u7684\u7b54\u6848\u3002\u989d\u5916\u7684\u65f6\u95f4\u7528\u4e8e\u786e\u4fdd\u5168\u9762\u548c\u6df1\u601d\u719f\u8651\u7684\u89e3\u51b3\u65b9\u6848\u3002<\/p>\n

2.3 \u5185\u5b58\u548c\u4e0a\u4e0b\u6587\u5904\u7406<\/h3>\n

\u4e24\u79cd\u6a21\u578b\u90fd\u53ef\u4ee5\u5904\u7406\u591a\u8fbe 64,000 \u4e2a\u8f93\u5165\u6807\u8bb0\uff0c\u4f46 DeepSeek-R1 \u7279\u522b\u64c5\u957f\u5728\u957f\u65f6\u95f4\u4ea4\u4e92\u4e2d\u4fdd\u6301\u903b\u8f91\u548c\u4e0a\u4e0b\u6587\u3002\u8fd9\u4f7f\u5f97\u5b83\u9002\u5408\u4e8e\u9700\u8981\u5728\u957f\u65f6\u95f4\u5bf9\u8bdd\u6216\u590d\u6742\u9879\u76ee\u4e2d\u6301\u7eed\u63a8\u7406\u548c\u7406\u89e3\u7684\u4efb\u52a1\u3002<\/p>\n

2.4 \u6700\u9002\u5408 API \u7528\u6237<\/h3>\n

\u5bf9\u4e8e\u4f7f\u7528 API \u7684\u7528\u6237\uff0cDeepSeek-V3 \u63d0\u4f9b\u4e86\u66f4\u81ea\u7136\u3001\u66f4\u6d41\u7545\u7684\u4ea4\u4e92\u4f53\u9a8c\u3002\u5b83\u5728\u8bed\u8a00\u548c\u5bf9\u8bdd\u65b9\u9762\u7684\u4f18\u52bf\u4f7f\u7528\u6237\u4ea4\u4e92\u611f\u89c9\u6d41\u7545\u4e14\u5f15\u4eba\u5165\u80dc\u3002<\/p>\n

R1 \u7684\u54cd\u5e94\u65f6\u95f4\u5bf9\u8bb8\u591a\u5e94\u7528\u7a0b\u5e8f\u6765\u8bf4\u53ef\u80fd\u662f\u4e00\u4e2a\u95ee\u9898\uff0c\u56e0\u6b64\u6211\u5efa\u8bae\u4ec5\u5728\u7edd\u5bf9\u5fc5\u8981\u65f6\u4f7f\u7528\u5b83\u3002<\/p>\n

\u8bf7\u6ce8\u610f\uff0c\u4f7f\u7528 API \u65f6\u7684\u6a21\u578b\u540d\u79f0\u4e0d\u662f V3 \u548c R1\u3002 V3 \u6a21\u578b\u540d\u4e3a deepseek-chat\uff0c\u800c R1 \u6a21\u578b\u540d\u4e3a deepseek-reasoner\u3002<\/p>\n

2.5 \u5b9a\u4ef7\u5dee\u5f02<\/h3>\n

\u5728\u8003\u8651\u4f7f\u7528\u54ea\u79cd\u6a21\u578b\u65f6\uff0c\u503c\u5f97\u6ce8\u610f\u7684\u662f V3 \u6bd4 R1 \u4fbf\u5b9c\u3002\u867d\u7136\u672c\u535a\u5ba2\u4fa7\u91cd\u4e8e\u529f\u80fd\uff0c\u4f46\u91cd\u8981\u7684\u662f\u8981\u6743\u8861\u4e0e\u6bcf\u79cd\u6a21\u578b\u76f8\u5173\u7684\u6210\u672c\u4ee5\u53ca\u6211\u4eec\u7684\u5177\u4f53\u9700\u6c42\u548c\u9884\u7b97\u3002\u6709\u5173\u6210\u672c\u7684\u66f4\u591a\u8be6\u7ec6\u4fe1\u606f\uff0c\u8bf7\u67e5\u770b\u5176\u3002<\/p>\n

3\u3001DeepSeek-R1 vs. V3\uff1aDeepSeek Chat\u793a\u4f8b<\/h2>\n

3.1 \u95ee\u9898\u89e3\u51b3\u548c\u903b\u8f91\u4efb\u52a1<\/h3>\n

\u8ba9\u6211\u4eec\u901a\u8fc7\u63d0\u51fa\u4ee5\u4e0b\u95ee\u9898\u6765\u6bd4\u8f83\u4e24\u79cd\u6a21\u578b\u7684\u63a8\u7406\u80fd\u529b\uff1a<\/p>\n

\"Use the digits [0-9] to make three numbers: x,y,z so that x+y=z\"<\/code><\/pre>\n
\n  \u7ffb\u8bd1\uff1a\u201c\u4f7f\u7528\u6570\u5b57 [0-9] \u7ec4\u6210\u4e09\u4e2a\u6570\u5b57\uff1ax\u3001y\u3001z\uff0c\u4f7f\u5f97 x+y=z\u201d\n<\/p><\/blockquote>\n
\u4f8b\u5982\uff0c\u4e00\u4e2a\u53ef\u80fd\u7684\u89e3\u51b3\u65b9\u6848\u662f\uff1ax = 26\u3001y = 4987 \u548c z = 5013\u3002\u5b83\u4f7f\u7528\u6240\u6709\u6570\u5b57 0-9 \u4e14 x + y = z\u3002<\/p>\n
\u5f53\u6211\u4eec\u5411 V3 \u63d0\u51fa\u8fd9\u4e2a\u95ee\u9898\u65f6\uff0c\u5b83\u7acb\u5373\u5f00\u59cb\u7ed9\u51fa\u4e00\u4e2a\u5197\u957f\u7684\u7b54\u6848\uff0c\u5e76\u6700\u7ec8\u5f97\u51fa\u4e86\u6ca1\u6709\u89e3\u51b3\u65b9\u6848\u7684\u9519\u8bef\u7ed3\u8bba\uff1a<\/p>\n
\u53e6\u4e00\u65b9\u9762\uff0cR1 \u7ecf\u8fc7\u5927\u7ea6 5 \u5206\u949f\u7684\u63a8\u7406\u540e\u5c31\u80fd\u627e\u5230\u89e3\u51b3\u65b9\u6848\uff1a<\/p>\n
\u8fd9\u8868\u660e R1 \u66f4\u9002\u5408\u9700\u8981\u6570\u5b66\u63a8\u7406\u7684\u95ee\u9898\uff0c\u56e0\u4e3a\u50cf V3 \u8fd9\u6837\u7684\u4e0b\u4e00\u4e2a\u5355\u8bcd\u9884\u6d4b\u4e0d\u592a\u53ef\u80fd\u8d70\u4e0a\u6b63\u786e\u7684\u9053\u8def\uff0c\u9664\u975e\u5728\u6a21\u578b\u8bad\u7ec3\u671f\u95f4\u4f7f\u7528\u4e86\u8bb8\u591a\u7c7b\u4f3c\u7684\u95ee\u9898\u3002<\/p>\n
3.2 \u521b\u4f5c\u5199\u4f5c<\/h3>\n
\u73b0\u5728\uff0c\u8ba9\u6211\u4eec\u5173\u6ce8\u521b\u610f\u5199\u4f5c\u3002\u8ba9\u6211\u4eec\u8981\u6c42\u4e24\u4e2a\u6a21\u578b\u5199\u4e00\u7bc7\u5173\u4e8e\u4eba\u7fa4\u4e2d\u5b64\u72ec\u7684\u5fae\u578b\u5c0f\u8bf4\u3002<\/p>\n
\"Write a microfiction story about loneliness in a crowd\"<\/code><\/pre>\n
\n  \u7ffb\u8bd1\uff1a\u201c\u5199\u4e00\u7bc7\u5173\u4e8e\u4eba\u7fa4\u4e2d\u5b64\u72ec\u7684\u5fae\u578b\u5c0f\u8bf4\u201d\n<\/p><\/blockquote>\n
\u8fd9\u662f V3 \u7684\u8f93\u51fa\uff1a<\/p>\n
\u6211\u4eec\u7acb\u5373\u5f97\u5230\u4e86\u4e00\u4e2a\u7b26\u5408\u4e3b\u9898\u7684\u6545\u4e8b\u3002\u6211\u4eec\u53ef\u80fd\u559c\u6b22\u6216\u4e0d\u559c\u6b22\uff0c\u8fd9\u662f\u4e3b\u89c2\u7684\uff0c\u4f46\u7b54\u6848\u4e0e\u6211\u4eec\u8981\u6c42\u7684\u4e00\u81f4\u3002<\/p>\n
\u5728\u4f7f\u7528\u63a8\u7406\u65f6\uff0c\u6a21\u578b\u63a8\u7406\u6765\u521b\u4f5c\u6545\u4e8b\u3002\u6211\u4eec\u4e0d\u4f1a\u5728\u8fd9\u91cc\u5c55\u793a\u6240\u6709\u7684\u7ec6\u8282\uff0c\u4f46\u5b83\u5c06\u4efb\u52a1\u5206\u89e3\u4e3a\u5982\u4e0b\u6b65\u9aa4\uff1a<\/p>\n
    \n
  • \u9996\u5148\uff0c\u6211\u5e94\u8be5\u8bbe\u7f6e\u573a\u666f\u2026\u2026<\/li>\n
  • \u63a5\u4e0b\u6765\uff0c\u611f\u5b98\u7ec6\u8282\u2026\u2026<\/li>\n
  • \u6211\u9700\u8981\u5c55\u793a\u4ed6\u4eec\u7684\u5185\u90e8\u72b6\u6001\u2026\u2026<\/li>\n
  • \u4ee5\u4e00\u4e2a\u51c4\u7f8e\u7684\u56fe\u50cf\u7ed3\u675f\u2026\u2026<\/li>\n
  • \u8ba9\u6211\u68c0\u67e5\u4e00\u4e0b\u6211\u662f\u5426\u6db5\u76d6\u4e86\u6240\u6709\u5143\u7d20\u2026\u2026<\/li>\n<\/ul>\n
    \u6211\u4eec\u53ef\u4ee5\u770b\u5230\uff0c\u521b\u4f5c\u8fc7\u7a0b\u975e\u5e38\u7ed3\u6784\u5316\uff0c\u8fd9\u53ef\u80fd\u4f1a\u964d\u4f4e\u8f93\u51fa\u7684\u521b\u9020\u529b\u3002<\/p>\n
    \u6211\u8ba4\u4e3a\uff0c\u53ea\u6709\u5f53\u6211\u4eec\u5bf9\u63a8\u7406\u8fc7\u7a0b\u611f\u5174\u8da3\u65f6\uff0c\u6211\u4eec\u624d\u5e94\u8be5\u5c06 R1 \u7528\u4e8e\u8fd9\u79cd\u4efb\u52a1\uff0c\u56e0\u4e3a\u6211\u4eec\u60f3\u8981\u7684\u8f93\u51fa\u4e0d\u662f\u903b\u8f91\u601d\u7ef4\u8fc7\u7a0b\u7684\u7ed3\u679c\uff0c\u800c\u662f\u521b\u9020\u6027\u7684\u7ed3\u679c\u3002<\/p>\n
    3.3 \u7f16\u7801\u534f\u52a9<\/h3>\n
    \u5728\u8fd9\u4e2a\u7b2c\u4e09\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u8981\u6c42 DeepSeek \u5e2e\u52a9\u4fee\u590d\u4e00\u4e2a\u7565\u6709\u9519\u8bef\u7684 Python \u51fd\u6570\uff0c\u8be5\u51fd\u6570\u65e8\u5728\u89e3\u51b3\u4ee5\u4e0b\u95ee\u9898\uff1a<\/p>\n
    \"People participating in a city run had to write down their names when starting and ending the race. We know that exactly one person didn't finish the race. This Python function is trying to find out the name of that person but it doesn't work. Fix it.\"\n\n```\ndef find_person(names):\n  freq = {}\n  # Calculate the frequency of each name\n  for name in names:\n    if name not in freq:\n      freq[name] = 0\n      freq[name] += 1\n  # Find the name that appears only once\n  for name in names:\n    if freq[name] == 1:\n      return name\n  return None\n```<\/code><\/pre>\n
    \n  \u7ffb\u8bd1\uff1a\u201c\u53c2\u52a0\u57ce\u5e02\u8dd1\u6b65\u7684\u4eba\u5fc5\u987b\u5728\u5f00\u59cb\u548c\u7ed3\u675f\u6bd4\u8d5b\u65f6\u5199\u4e0b\u4ed6\u4eec\u7684\u540d\u5b57\u3002\u6211\u4eec\u77e5\u9053\u53ea\u6709\u4e00\u4e2a\u4eba\u6ca1\u6709\u5b8c\u6210\u6bd4\u8d5b\u3002\u8fd9\u4e2a Python \u51fd\u6570\u8bd5\u56fe\u627e\u51fa\u90a3\u4e2a\u4eba\u7684\u540d\u5b57\uff0c\u4f46\u5b83\u4e0d\u8d77\u4f5c\u7528\u3002\u4fee\u590d\u5b83\u3002\u201d\n<\/p><\/blockquote>\n
    \u5728\u5c06\u5176\u53d1\u9001\u7ed9 AI \u4e4b\u524d\uff0c\u8ba9\u6211\u4eec\u5148\u4e86\u89e3\u4e00\u4e0b\u4ee3\u7801\u4e2d\u7684\u95ee\u9898\u3002<\/p>\n
    \u7531\u4e8e\u6bcf\u4e2a\u4eba\u5728\u5f00\u59cb\u548c\u7ed3\u675f\u6bd4\u8d5b\u65f6\u90fd\u4f1a\u5199\u4e0b\u81ea\u5df1\u7684\u540d\u5b57\uff0c\u56e0\u6b64\u6b64\u4ee3\u7801\u8bd5\u56fe\u901a\u8fc7\u67e5\u627e\u53ea\u51fa\u73b0\u4e00\u6b21\u7684\u540d\u5b57\u6765\u89e3\u51b3\u95ee\u9898\u3002\u6bcf\u4e2a\u5b8c\u6210\u6bd4\u8d5b\u7684\u4eba\u90fd\u4f1a\u5199\u4e24\u6b21\u81ea\u5df1\u7684\u540d\u5b57\uff0c\u800c\u6ca1\u6709\u5b8c\u6210\u6bd4\u8d5b\u7684\u4eba\u53ea\u4f1a\u5199\u4e00\u6b21\u3002\u7136\u800c\uff0c\u6b64\u4ee3\u7801\u9519\u8bef\u5730\u5047\u8bbe\u6240\u6709\u540d\u5b57\u90fd\u662f\u4e0d\u540c\u7684\u3002<\/p>\n
    \u6b63\u786e\u7684\u7b54\u6848\u4e0d\u662f\u9891\u7387\u7b49\u4e8e 1 \u7684\u540d\u5b57\uff0c\u800c\u662f\u9891\u7387\u4e3a\u5947\u6570\u7684\u540d\u5b57\u3002\u56e0\u6b64\uff0c\u89e3\u51b3\u65b9\u6848\u662f\u5c06\u7b2c\u4e8c\u4e2a for \u5faa\u73af\u68c0\u67e5 if freq[name] == 1:<\/code> \u66ff\u6362\u4e3a if freq[name] % 2 == 1<\/code>\uff0c\u4ee5\u67e5\u627e\u9891\u7387\u5947\u6570\u7684\u540d\u79f0\u3002<\/p>\n
    \u8ba9\u6211\u4eec\u770b\u770b\u8fd9\u4e24\u4e2a\u6a21\u578b\u5982\u4f55\u89e3\u51b3\u8fd9\u4e2a\u95ee\u9898\u3002\u6211\u4eec\u63d0\u4f9b\u4e86\u4ee5\u4e0b\u63d0\u793a\uff1a<\/p>\n
    \u8fd9\u662f V3 \u7684\u54cd\u5e94\uff1a<\/p>\n
    V3 \u6a21\u578b\u65e0\u6cd5\u627e\u5230\u6b63\u786e\u7b54\u6848\u3002\u5b83\u4e0d\u4ec5\u901a\u8fc7\u5f15\u5165\u4e24\u4e2a\u8f93\u5165\u5217\u8868\u6765\u66f4\u6539\u95ee\u9898\u53c2\u6570\uff0c\u800c\u4e14\u5373\u4f7f\u6211\u4eec\u6709\u4e24\u4e2a\u4e0d\u540c\u7684\u5217\u8868\uff0c\u63d0\u4f9b\u7684\u89e3\u51b3\u65b9\u6848\u4e5f\u65e0\u6cd5\u5de5\u4f5c\u3002<\/p>\n
    \u76f8\u6bd4\u4e4b\u4e0b\uff0cR1 \u53ef\u4ee5\u627e\u5230\u4ee3\u7801\u4e2d\u7684\u95ee\u9898\uff0c\u5373\u4f7f\u5b83\u7684\u89e3\u51b3\u65b9\u6848\u66f4\u6539\u4e86\u4ee3\u7801\u800c\u4e0d\u662f\u4fee\u590d\u63d0\u4f9b\u7684\u4ee3\u7801\uff1a<\/p>\n
    \u8be5\u6a21\u578b\u5728\u5bfb\u627e\u7b54\u6848\u65f6\u76f8\u5f53\u6162\u3002\u6211\u4eec\u770b\u5230\u5b83\u63a8\u7406\u4e86\u5c06\u8fd1\u516b\u5206\u949f\u3002\u7a81\u51fa\u663e\u793a\u7684\u90e8\u5206\u663e\u793a\u4e86\u6a21\u578b\u4f55\u65f6\u610f\u8bc6\u5230\u4ee3\u7801\u51fa\u4e86\u4ec0\u4e48\u95ee\u9898\u3002<\/p>\n
    4\u3001\u4f55\u65f6\u9009\u62e9 DeepSeek R1 \u6216 V3<\/h2>\n
    \u5728 DeepSeek-R1 \u548c DeepSeek-V3 \u4e4b\u95f4\u9009\u62e9\u6b63\u786e\u7684\u6a21\u578b\u53d6\u51b3\u4e8e\u4f60\u5e0c\u671b\u901a\u8fc7\u6211\u4eec\u7684\u4efb\u52a1\u6216\u9879\u76ee\u5b9e\u73b0\u4ec0\u4e48\u76ee\u6807\u3002<\/p>\n
    \u5bf9\u4e8e\u5927\u591a\u6570\u4efb\u52a1\uff0c\u6211\u4e00\u822c\u63a8\u8350\u7684\u5de5\u4f5c\u6d41\u7a0b\u662f\u4f7f\u7528 V3\uff0c\u5982\u679c\u9677\u5165 V3 \u65e0\u6cd5\u627e\u5230\u7b54\u6848\u7684\u5faa\u73af\uff0c\u5219\u5207\u6362\u5230 R1\u3002\u4f46\u662f\uff0c\u6b64\u5de5\u4f5c\u6d41\u7a0b\u5047\u8bbe\u6211\u4eec\u53ef\u4ee5\u786e\u5b9a\u6211\u4eec\u5f97\u5230\u7684\u7b54\u6848\u662f\u5426\u6b63\u786e\u3002\u6839\u636e\u95ee\u9898\u7684\u4e0d\u540c\uff0c\u6211\u4eec\u53ef\u80fd\u5e76\u4e0d\u603b\u662f\u80fd\u591f\u505a\u51fa\u8fd9\u79cd\u533a\u5206\u3002<\/p>\n
    \u4f8b\u5982\uff0c\u5728\u7f16\u5199\u4e00\u4e2a\u603b\u7ed3\u4e00\u4e9b\u6570\u636e\u7684\u7b80\u5355\u811a\u672c\u65f6\uff0c\u6211\u4eec\u53ef\u4ee5\u8fd0\u884c\u4ee3\u7801\u5e76\u67e5\u770b\u5b83\u662f\u5426\u5728\u6267\u884c\u6211\u4eec\u60f3\u8981\u7684\u64cd\u4f5c\u3002\u4f46\u662f\uff0c\u5982\u679c\u6211\u4eec\u6b63\u5728\u6784\u5efa\u4e00\u4e2a\u590d\u6742\u7684\u7b97\u6cd5\uff0c\u9a8c\u8bc1\u4ee3\u7801\u662f\u5426\u6b63\u786e\u5c31\u4e0d\u90a3\u4e48\u7b80\u5355\u4e86\u3002<\/p>\n
    \u56e0\u6b64\uff0c\u5728\u4e24\u79cd\u6a21\u578b\u4e4b\u95f4\u8fdb\u884c\u9009\u62e9\u65f6\uff0c\u4ecd\u7136\u9700\u8981\u4e00\u4e9b\u6307\u5bfc\u65b9\u9488\u3002\u4ee5\u4e0b\u662f\u4f55\u65f6\u9009\u62e9\u5176\u4e2d\u4e00\u4e2a\u7684\u6307\u5357\uff1a<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n
    \u4efb\u52a1<\/th>\n\u6a21\u578b<\/th>\n<\/tr>\n<\/thead>\n
    \u5199\u4f5c\u3001\u5185\u5bb9\u521b\u5efa\u3001\u7ffb\u8bd1<\/td>\nV3<\/td>\n<\/tr>\n
    \u53ef\u4ee5\u8bc4\u4f30\u8f93\u51fa\u8d28\u91cf\u7684\u4efb\u52a1<\/td>\nV3<\/td>\n<\/tr>\n
    \u4e00\u822c\u7f16\u7801\u95ee\u9898<\/td>\nV3<\/td>\n<\/tr>\n
    AI \u52a9\u624b<\/td>\nV3<\/td>\n<\/tr>\n
    \u7814\u7a76<\/td>\nR1<\/td>\n<\/tr>\n
    \u590d\u6742\u7684\u6570\u5b66\u3001\u7f16\u7801\u6216\u903b\u8f91\u95ee\u9898<\/td>\nR1<\/td>\n<\/tr>\n
    \u4e3a\u89e3\u51b3\u5355\u4e2a\u95ee\u9898\u800c\u8fdb\u884c\u7684\u957f\u65f6\u95f4\u8fed\u4ee3\u5bf9\u8bdd<\/td>\nR1<\/td>\n<\/tr>\n
    \u6709\u5174\u8da3\u4e86\u89e3\u5f97\u51fa\u7b54\u6848\u7684\u601d\u7ef4\u8fc7\u7a0b<\/td>\nR1<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
    5\u3001\u7ed3\u675f\u8bed<\/h2>\n
    DeepSeek V3 \u975e\u5e38\u9002\u5408\u65e5\u5e38\u4efb\u52a1\uff0c\u5982\u5199\u4f5c\u3001\u5185\u5bb9\u521b\u5efa\u548c\u5feb\u901f\u7f16\u7801\u95ee\u9898\uff0c\u4ee5\u53ca\u6784\u5efa AI \u52a9\u624b\uff0c\u5176\u4e2d\u81ea\u7136\u3001\u6d41\u7545\u7684\u5bf9\u8bdd\u662f\u5173\u952e\u3002\u5b83\u4e5f\u975e\u5e38\u9002\u5408\u53ef\u4ee5\u5feb\u901f\u8bc4\u4f30\u8f93\u51fa\u8d28\u91cf\u7684\u4efb\u52a1\u3002<\/p>\n
    \u4f46\u662f\uff0c\u5bf9\u4e8e\u9700\u8981\u6df1\u5ea6\u63a8\u7406\u7684\u590d\u6742\u6311\u6218\uff0c\u4f8b\u5982\u7814\u7a76\u3001\u590d\u6742\u7684\u7f16\u7801\u6216\u6570\u5b66\u95ee\u9898\uff0c\u6216\u6269\u5c55\u7684\u89e3\u51b3\u95ee\u9898\u5bf9\u8bdd\uff0cDeepSeek R1 \u662f\u66f4\u597d\u7684\u9009\u62e9\u3002<\/p>\n
    \n","protected":false},"excerpt":{"rendered":"
    \u5728\u624b\u673a\u6216\u684c\u9762\u4e0a\u4f7f\u7528 DeepSeek \u5e94\u7528\u7a0b\u5e8f\u65f6\uff0c\u6211\u4eec\u53ef\u80fd\u4f1a\u4e0d\u786e\u5b9a\u4f55\u65f6\u9009\u62e9 R1\uff08\u4e5f\u79f0\u4e3a DeepThink\uff09\uff0c\u800c\u4e0d\u662f\u65e5\u5e38\u4efb\u52a1\u7684\u9ed8\u8ba4 V3 \u6a21\u578b\u3002 \u5bf9\u4e8e\u5f00\u53d1\u4eba\u5458\u6765\u8bf4\uff0c\u6311\u6218\u6709\u70b9\u4e0d\u540c\u3002\u5f53\u901a\u8fc7\u5176 API \u96c6\u6210 DeepSeek \u65f6\uff0c\u6311\u6218\u5728\u4e8e\u627e\u51fa\u54ea\u79cd\u6a21\u578b\u66f4\u7b26\u5408\u6211\u4eec\u7684\u9879\u76ee\u8981\u6c42\u5e76\u589e\u5f3a\u529f\u80fd\u3002 \u5728\u6b64\u535a\u5ba2\u4e2d\uff0c\u6211\u5c06\u4ecb\u7ecd\u8fd9\u4e24\u79cd\u6a21\u578b\u7684\u5173\u952e\u65b9\u9762\uff0c\u4ee5\u5e2e\u52a9\u4f60\u66f4\u8f7b\u677e\u5730\u505a\u51fa\u8fd9\u4e9b\u9009\u62e9\u3002\u6211\u5c06\u63d0\u4f9b\u793a\u4f8b\u6765\u8bf4\u660e\u6bcf\u4e2a\u6a21\u578b\u5728\u4e0d\u540c\u60c5\u51b5\u4e0b\u7684\u884c\u4e3a\u548c\u6027\u80fd\u3002\u6211\u8fd8\u4f1a\u4e3a\u4f60\u63d0\u4f9b\u4e00\u4e2a\u51b3\u7b56\u6307\u5357\uff0c\u53ef\u4ee5\u4f7f\u7528\u5b83\u6765\u5728 DeepSeek-R1 \u548c DeepSeek-V3 \u4e4b\u95f4\u8fdb\u884c\u9009\u62e9\u3002 1\u3001DeepSeek-V3 \u548c DeepSeek-R1 DeepSeek \u662f\u4e00\u5bb6\u4e2d\u56fd AI \u521d\u521b\u516c\u53f8\uff0c\u5b83\u4ee5\u6bd4 OpenAI \u7684 o1 \u4f4e\u5f97\u591a\u7684\u6210\u672c\u5f00\u53d1\u4e86 DeepSeek-R1\uff0c\u5f15\u8d77\u4e86\u56fd\u9645\u5173\u6ce8\u3002\u5c31\u50cf OpenAI \u6709\u6211\u4eec\u90fd\u77e5\u9053\u7684 ChatGPT \u5e94\u7528\u7a0b\u5e8f\u4e00\u6837\uff0cDeepSeek \u4e5f\u6709\u7c7b\u4f3c\u7684\u804a\u5929\u673a\u5668\u4eba\uff0c\u5b83\u6709\u4e24\u4e2a\u6a21\u578b\uff1aDeepSeek-V3 \u548c DeepSeek-R1\u3002 1.1 \u4ec0\u4e48\u662f DeepSeek-V3\uff1f DeepSeek-V3 \u662f\u6211\u4eec\u4e0e DeepSeek \u5e94\u7528\u7a0b\u5e8f\u4ea4\u4e92\u65f6\u4f7f\u7528\u7684\u9ed8\u8ba4\u6a21\u578b\u3002\u5b83\u662f\u4e00\u79cd\u591a\u529f\u80fd\u7684\u5927\u578b\u8bed\u8a00\u6a21\u578b (LLM)\uff0c\u662f\u4e00\u79cd\u53ef\u4ee5\u5904\u7406\u5404\u79cd\u4efb\u52a1\u7684\u901a\u7528\u5de5\u5177\u3002 \u8be5\u6a21\u578b\u4e0e\u5176\u4ed6\u77e5\u540d\u8bed\u8a00\u6a21\u578b\uff08\u5982 OpenAI \u7684 GPT-4o\uff09\u7ade\u4e89\u3002 DeepSeek-V3 \u7684\u4e3b\u8981\u529f\u80fd\u4e4b\u4e00\u662f\u5b83\u4f7f\u7528\u4e86\u6df7\u5408\u4e13\u5bb6 (MoE) \u65b9\u6cd5\u3002\u8fd9\u79cd\u65b9\u6cd5\u5141\u8bb8\u6a21\u578b\u4ece\u4e0d\u540c\u7684\u201c\u4e13\u5bb6\u201d\u4e2d\u8fdb\u884c\u9009\u62e9\u6765\u6267\u884c\u7279\u5b9a\u4efb\u52a1\u3002\u5728\u4f60\u5411\u6a21\u578b\u63d0\u4f9b\u63d0\u793a\u540e\uff0c\u53ea\u6709\u6a21\u578b\u4e2d\u6700\u76f8\u5173\u7684\u90e8\u5206\u624d\u4f1a\u9488\u5bf9\u4efb\u4f55\u7ed9\u5b9a\u4efb\u52a1\u5904\u4e8e\u6d3b\u52a8\u72b6\u6001\uff0c\u4ece\u800c\u8282\u7701\u8ba1\u7b97\u8d44\u6e90\u5e76\u63d0\u4f9b\u7cbe\u786e\u7684\u7ed3\u679c\u3002 \u672c\u8d28\u4e0a\uff0cDeepSeek-V3 \u662f\u6211\u4eec\u9700\u8981 LLM […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[13],"tags":[],"class_list":["post-53788","post","type-post","status-publish","format-standard","hentry","category-ai"],"_links":{"self":[{"href":"https:\/\/fwq.ai\/blog\/wp-json\/wp\/v2\/posts\/53788","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fwq.ai\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/fwq.ai\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/fwq.ai\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/fwq.ai\/blog\/wp-json\/wp\/v2\/comments?post=53788"}],"version-history":[{"count":0,"href":"https:\/\/fwq.ai\/blog\/wp-json\/wp\/v2\/posts\/53788\/revisions"}],"wp:attachment":[{"href":"https:\/\/fwq.ai\/blog\/wp-json\/wp\/v2\/media?parent=53788"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/fwq.ai\/blog\/wp-json\/wp\/v2\/categories?post=53788"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/fwq.ai\/blog\/wp-json\/wp\/v2\/tags?post=53788"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}