国产大模型，贵到用不起 - 投资界
High confidence: full text extraction produced 8760 characters.
æè¿è¿ä¸ä¸ªæï¼ç¡ è°·AIçå¯è°å¤§æé¢åºï¼GPT-5.6ãfable5ãGrok-4.5æ¥è¿é®ä¸ã
ä¸æ¶é´ï¼å½å å¼åè ãç¨åºå群ä½çº·çº·æ²¸è ¾ï¼åç§å ³äºGPT-5.6çå沿模åç讨论ãæµè¯æ¤èµ·å½¼ä¼ã
尽管è¿ä¸ªæä¹æKimi-K3è¿æ ·çæ°éå½äº§æ¨¡åé®ä¸ï¼ä½æ»çæ¥çï¼å½äº§å¤§æ¨¡åå½åå¨ä¸çº¿å¼åè 群ä½ä¸ï¼å§ç»å¤äºä¸ç§âå«å¥½ä¸å«åº§âçç¶æã
æ ¹æ®ææ°æ°æ®ï¼å¨ä½¿ç¨éæ¹é¢ï¼OpenRouter 6æ22æ¥è³28æ¥æ°æ®æ¾ç¤ºï¼ä¸å½æ¨¡åå¨è°ç¨éè¾¾20.39ä¸äº¿Tokenï¼ç¾å½æ¨¡å为4.25ä¸äº¿ï¼åè å·²è¿ç»ä¹å¨é¢ å ã
ä½ä¸ä¹å½¢æåå·®çæ¯ï¼å½å AIå å°é¾å¨è¥æ¶æ¹é¢ç表ç°ï¼å´ä¸æä¹å°½å¦äººæï¼å ¬å¼è´¢æ¥æ¾ç¤ºï¼æºè°±2025å¹´è¥æ¶7.24亿å ï¼äºæå´è¾¾å°äº30亿å 以ä¸ï¼MiniMax为7900ä¸ç¾å ï¼çº¦5.7亿å ï¼ï¼æä¹æé¢çå ¶ä½å家å°æªæ«é²å®æ´å¹´è¥æ¶ã
å³ä½¿å¤´é¨ä¼ä¸ï¼æ¶å ¥ä»åçå¨æ°äº¿å é级ã
ç¸è¾ä¹ä¸ï¼ä½ä¸ºAI头é¨ä¼ä¸çAnthropicï¼å ¶ARRå·²ç»è¾¾å°äºâ约 600 亿â700 亿ç¾å âï¼ä¸»è¦é©±å¨å为 B2B API å AI Coding åºæ¯ ãââ
è¿ä¸ç±å¾å¼åºäºä¸ä¸ªå°éçé®é¢ï¼
为ä½å·ç§°â便å®âãâè°ç¨é大âçå½äº§æ¨¡åï¼è³ä»ä»ç¶å«å¥½ä¸å«åº§ï¼å¨AI Codingçé«ä»·å¼çåºæ¯ä¸ï¼è³ä»ä»é¾ä»¥ä¸å½å¤é¡¶ å°æ¨¡åæ£é¢æè¡¡ï¼
æè¿ï¼ç¬è å¨ä¸å¤ä¸ªä¸çº¿å¼åè æ·±å ¥æ¢è®¨æ¤é®é¢åï¼åç°äºä¸ä¸ªå¾åç´è§çäºå®ï¼
å½äº§å¤§æ¨¡åï¼æç§ç¨åº¦ä¸å ¶å®é常âè´µâã
ä¹æ£æ¯è¿æ ·çâè´µâï¼è®©æäºæ§è½ä¸ææçªç ´çå½äº§æ¨¡åï¼é¾ä»¥çæ£å±ç°èªå·±çéªå ä¹å¤ã
01
âä½ä»·âçå¹»è§
æèµ·å½äº§å¤§æ¨¡åï¼å¾å¤äººæ³å°çé½æ¯DeepSeekãKimiãæºè°±è¿æ ·ç代表ï¼èä¸ä¹å ³èæç´§å¯çæ ç¾ä¹ä¸ï¼å°±æ¯ä½ä»·ã便å®ï¼
å¦æä» ä»Tokenåä»·ä¸æ¥çï¼å½äº§æ¨¡åä¸GPT-5.6ãfable5ç顶 å°æ¨¡åç¸æ¯ï¼ç¡®å®æçä¸å°çä¼å¿ï¼æ¿GLM-5.2ä¸DeepSeek V4举ä¾ï¼GLM-5.2æ¯ç¾ä¸Tokenè¾å ¥/è¾åºä¸º1.4/4.4ç¾å ï¼DeepSeek-V4-Pro为0.435/0.87ç¾å ï¼GPT-5.6 Sol为5/30ç¾å ï¼Claude Fable 5å为10/50ç¾å ã
ä½å®é ä¸ï¼è¿ç§âä½ä»·âçå°è±¡ï¼æ¯ä¸ç§åå¤äººé¿ä¹ ç误åºã
å°¤å ¶å¨é«å¼ºåº¦çç¼ç¨åºæ¯ä¸ï¼å½äº§å¤§æ¨¡åçåä»·ä¸ä½æ²¡æä¼å¿ï¼çè³åèè¿æ¯æå¥é¤çGPTãClaudeè´µåºå¥½å åã
æå ï¼ååï¼æ¯AIé¢åçä¸ä½ç§ç å·¥ä½è ï¼ç±äºå·¥ä½éè¦ï¼æ¯å¤©é½è¦å¨AI Codingçåºæ¯ä¸æ¶èå å亿Tokenï¼å¨è¿æ ·çæ¶èéä¸ï¼å¦æç¨å½äº§å¤§æ¨¡åï¼ä¾å¦GLM-5.2ï¼æ´ä½çææ¬å°±ä¼åå¾é常ææã
æå æåºçTokenè´¦å
å¨äº¤è°å½å¤©ï¼æå æ¶èçTokenå·²ç»æ¥è¿40亿ï¼è¿éå¯ä»¥ç¨å¾®æ¢ç®ä¸ä¸ï¼åè¥ä»ç¨çæ¯GLM-5.2ï¼ææ¬å¤§è´æ¯ï¼ï¼8639ä¸æ®éè¾å ¥Ã1.4ç¾å ï¼33.80亿ç¼åè¾å ¥Ã0.26ç¾å ï¼1906ä¸è¾åºåæ¨çÃ4.4ç¾å ï¼çº¦ä¸º1084ç¾å ï¼æ1ç¾å å 7.2å ä¼°ç®ï¼çº¦å人æ°å¸7800å ã
è让æå å¦æ¤æ¾å¼æèç§Tokençåå ï¼å°±æ¯å 为ä»å¼äºGPTçCodeXå¥é¤ï¼
è¿éç®å解éä¸ä¸CodeXå¥é¤çå ·ä½å½¢å¼ï¼ç®åæ¥è¯´ï¼å®æ¯é对ç¼ç¨åºæ¯æ¨åºçä¸æ¬¾å¥é¤ï¼ä¸¥æ ¼è¯´æ¥ï¼å®å¹¶éçæ£ä¸ééï¼èæ¯æCodexçº³å ¥ChatGPT订é ï¼Plus为20ç¾å /æï¼Proä»100ç¾å /æèµ·ï¼é¢åº¦çº¦ä¸ºPlusç5åæ20åï¼è§¦é¡¶åå¯è´ä¹°Credits继ç»ä½¿ç¨ã
åæ ·å°ï¼AnthropicçClaude Codeä¹æä¸å¥ç±»ä¼¼çå¥é¤ï¼å ·ä½å½¢å¼æ¯ï¼Claude Pro为20ç¾å /æï¼Max 5xåMax 20xåå«ä¸º100ã200ç¾å /æï¼Claudeç½é¡µç«¯ä¸Claude Codeå ±äº«æ¯5å°æ¶åæ¯å¨é¢åº¦ï¼è§¦é¡¶åå¯æAPIä»·æ ¼ç»ç¨ã
å¨è¿æ ·çå¥é¤å åºä¸ï¼å·¨å¤§çTokenææ¬åè被æ¹å¹³äºã
对äºå½å çå¼åè æ¥è¯´ï¼å¨é«å¼ºåº¦AI Codingåºæ¯ä¸ï¼æ¯å¤©æ¶èå å亿ãä¸ç¾äº¿Tokenï¼æ¯ä¸ä¸ªå¾æ®éç常æã
å°åä¹æ¯ä¸ä¸ªAI Codingçé度å¼åè ï¼ä»å¨Codexä¸å®é ä¸å¨çè±è´¹æ¯ 300å 人æ°å¸ï¼èè¿è¿æ¯å¨æ²¡æå ¬å¸æ¥éçæ åµä¸ã
ä½å³ä½¿æ¯è¿æ ·çææ¬ï¼å¯¹æ¯å½äº§æ¨¡åï¼ä¾ç¶æ¾å¾â便å®âäºå¾å¤ï¼å 为åæ ·300å ï¼æ ¹æ¬ä¹°ä¸æ¥GLM 5.2 åçæ°éçtokenã
å°åçTokenè´¦å
å¨å½å å¼åè çæ¥ï¼GPTç codex å¥é¤ï¼æ¯ç®åè½ä¹°å°çæ 便å®ç tokenï¼æ¯æºè°±ãkimi 便å®å¥½å åï¼è¿å¯è½åç´è§ï¼ä½æ¯äºå®å°±æ¯å½äº§æ¨¡åå¨é«å¼ºåº¦çCodingåºæ¯ä¸ï¼æ§ä»·æ¯å®å ¨æ æ³ä¸ä¹å¯¹æ¯ã
è¿ééè¦è¯´æçæ¯ï¼codex并éçæ£æä¹çä¸ééå¥é¤ï¼å¦æææé«æ¡£200ç¾å ç®ï¼ä¸å¨å¤§æ¦æ20亿é¢åº¦ã
ä½æ¯ï¼ç±äºé¨åå¼åè ä¼éè¿âä¸è½¬ç«âçæ段使ç¨æ¨¡åï¼æäºæ¨¡åçå®é ææ¬ï¼ä¼æ¯å®æ¹æ¸ éä½å¾å¤ï¼è¿å°±å¯¼è´ææ¶è½ä»¥æ´ä½çä»·æ ¼ï¼ä¹°å°æ´å¤Tokenï¼äºæ¯ææäº300åè½ä¹°120亿Tokençç°è±¡ã
å®é ä¸ï¼å½äº§æ¨¡åå¨å®ä»·æ¹é¢ï¼ä¸æ¯æ²¡ææ¨åºè¿ç±»ä¼¼çå¥é¤ï¼ä½å¦æç»ç©¶èµ·æ¥ï¼è¿æ ·çå¥é¤ï¼å¨éé¢æ¹é¢ï¼æ¯èµ·codexèè¨ï¼åå¨çå¾å¤§çéå¶ã
02
å½äº§å®ä»·ä¹ç
å½äº§å¤§æ¨¡åå¨å®ä»·æ¹é¢çç°ç¶ï¼å¯ä»¥ç¨ä¸å¥è¯æ¥å½¢å®¹ï¼
âä½ ä»¥ä¸ºä½ è¿çæ¯èªå©é¤çåºåï¼ä½æååç°ï¼å家è¿æ¯ææ¤è®¡è´¹ãâ
举ä¾æ¥è¯´ï¼å¨å¥é¤æ¹é¢ï¼å½å çKimiãGLM-5.2ï¼Kimi Codeå为49ã99ã199ã699å åæ¡£ï¼é¢åº¦æå¨å·æ°ï¼GLM Coding PlançLiteãProãMaxæ¯5å°æ¶åå«çº¦80ã400ã1600次Promptï¼æ¯å¨çº¦400ã2000ã8000次ï¼ä¸GLM-5.2é«å³°ææ3åæ£é¢ã
以é¿éçqoder为ä¾ï¼æ ¹æ®ç¬è ä¸ä½ä»äºB端客æ·è¿ç»´çæåé¿è¿·ï¼ååï¼ä»ç»ï¼âqoderæ¾ç»æ¯å½å æ 好ç¨çAI IDEï¼å¯æå®èªå·±çæ°å®ä»·ï¼æä¼å¿æ没äºãâ
å¨å¥é¤æ¹é¢ï¼Qoder CN个人ProãPro+åå«ä¸º59ã169å /æï¼å«2000ã6000 Creditsï¼Teamsä¸VPCçåå«ä¸º99ã199å /å¸ä½Â·æï¼åå«3000 Creditsï¼åè 50å¸èµ·å®ã
âä»å¹´ä»ä»¬å并çµç è°äºå®ä»·åï¼å°±æç¥éçï¼è³å°é½æäºä¸å家æ¾å¼ç»è´¹äºãâé¿è¿·åæ¥ä»ç»éã
æ ¹æ®é¿è¿·çæµè¯ï¼qoderä¼ä¸ç199çå¥é¤ï¼æä»çç¨æ³ï¼3天就没äºï¼å¦ææé度å¼åè çç¨æ³ï¼ä¼°è®¡ä¸å¤©å°±æ²¡äºã
åæ ·å°ï¼å¨å¥é¤æ¹é¢ï¼ä»¤å½å ç¨æ·æå°äºå³éæçï¼è¿æGLM-5.2ã
æç §å¼åè å°èç说æ³ï¼âGLM-5.2çå¥é¤çº¦çäºæ²¡æï¼å°±ç®æå¥é¤é«å³°æä»ç¶éæµï¼å¹¶ä¸è¿åå¨âèåºâç¨æ·çç°è±¡ãâ
æç §36æ°ªçæ¥éï¼GLM-5.2çå¥é¤ï¼ä¹åæ¯æ¼å±ä¼é¨ç¥¨è¿é¾æ¢ï¼æ¯å¤©æ©ä¸åç¹éåæ¶å·æ°ï¼è½ä¸è½ä¹°å°å ¨åæéã
2026 å¹´ 1 æï¼æºè°±åå ¬å说å 为ç®åç´§å¼ ï¼ææ¯æ¥å¯éå®éç´æ¥ç å°åæ¥ç 20%ï¼æ¯å¤©æ©ä¸ 10 ç¹æ¾ä¸ç¹é¢åº¦ï¼å åéå°±å®ç½ã
说å°åºï¼å½äº§æ¨¡åå¨å¥é¤ãä»·æ ¼ä¸çâåå¬âï¼æ¬è´¨ä¸æ¯ç®åä¸è¶³é æççªè¿«ã
æ ¹æ®2025å¹´ä¸å½ä¿¡éé¢ã工信é¨çæ°æ®ï¼ä¸å½ç°åçæ°æ®ä¸å¿æ¯449 个ï¼èç¾å½åæ¯5427 个ï¼å¨æ»ç®åæ¹é¢ï¼ä¸å½ä¸º1053 EFLOPSï¼èç¾å½åæ¯2400ã
ä½è¿éçå ³é®å¨äºï¼ç¾å½é£2,400 EFLOPS éï¼é«ç«¯ GPU å æ¯æé«ï¼ä¸å½çç®åéï¼å¤§éæ¯å为æè ¾ãå¯æ¦çºªçå½äº§è¯çï¼åå¡æ§è½ä¸ H100 ä»æ代差ã
为åºå¯¹æåé«æ¶¨çç®åéæ±ï¼2026 å¹´ 3 æå¼å§ï¼æºè°±ãé¿éäºãè ¾è®¯äºãç¾åº¦å¨Tokenä»·æ ¼ä¸éä½æ¶¨ä»·ï¼æ¶¨å¹ ä» 5% å° 460% ä¸çã
说ç½äºï¼å½äº§æ¨¡åçâä½ä»·âãâä»·æ ¼æâæ©å·²æ为äºè¿å»å¼ã
ä¸æ¯å½äº§æ¨¡åä¸æ³æbuffetï¼æ¯å 为ç®è¿è´¦ï¼ä»¥ç°å¨çç®åææ¬åäºæç¶æï¼æå æçäºæ¿åºå®æè´¹å»èµç¨æ·ä¸ä¼å·çï¼èæç §å½å å¼åè ãç¨åºåå¨è¾ä¸å¤©ç§å å亿ãçè³ä¸ç¾äº¿Tokençç¨æ³ï¼è¿æ ·çåä¸æ¨¡å¼ï¼å¾å¿«ä¼è¢«å»ç©¿ææ¬çº¿ã
èOpenAI å Anthropic çç¼ç¨å¥é¤ï¼åçæ¯é«åå¼ä¼ä¸å®¢æ·ââCursorãGitHub Copilot è¿ç§å¤§å®¢æ·ä¸å¹´è½ç» Anthropic è´¡ç® 14 亿ç¾å æ¶å ¥ã
å¨é«ç®åçå æä¸ï¼è¿å¥åä¸æ¨¡å¼ï¼æ¬è´¨æ¯ç¨"åºå®æè´¹"æ¢"é¿æé客"ï¼å®¢æ·è´¨éé«ï¼ARPU ä¹ç¸å¯¹è¾é«ï¼æå¾å¹³ææ¬ã
èç¸è¾ä¹ä¸ï¼å½å è½ç¶æé¿éè¿æ ·çäºå·¨å¤´ï¼ä½é®é¢æ¯ï¼é¿éäºFY2026 è°æ´å EBITA Margin åªæ 7.5%ï¼äºæºè½éå¢è½ç¶æ¶å ¥å¢é¿å¿«ï¼ä½å©æ¶¦å¹¶ä¸ä¸°åã
并ä¸ï¼ä¼ ç»äºååæ¢å"ä¸ééäºæå¡å¨"ï¼æ¯å 为ç¨æ·ä¸æ¯ 24 å°æ¶æ»¡è´è½½ãä¸å° ECSï¼å®é CPU å©ç¨çå¯è½åªæ 10%ï¼äºååå¯ä»¥æä¸å°ç©çæºè¶ åç»å个人ã
ä½å¤§æ¨¡åæ¨çä¸ä¸æ ·ï¼æ¥ä¸ä¸ªtokenï¼å°±è¦å®æå®ç§ä¸æ¬¡ GPU ç®åãç¨æ·å¦æ 24 å°æ¶ä¸é´æå·ï¼äºååç HBM æ¾åãGPU æ ¸å¿ãçµè´¹å ¨æ¯åæ§æ¯åºï¼æ²¡æä»»ä½è¶ å空é´ã
03
模åçâé»éä¸è§â
é¤äºä»·æ ¼æ¹é¢çå ç´ å¤ï¼å¯¹å½å å¼åè æ¥è¯´ï¼éæ©CodexãCalude Codeçåå ï¼æ çæ¯å ¶å¼ºå¤§çç¼ç¨æ§è½ã
ç¶èï¼å¨æ§è½æ¹é¢ï¼å½äº§æ¨¡åä¹å¹¶éåæäºå»æ¿å°è±¡ä¸æ³çé£æ ·ã
å¨ä¸å¤ä¸ªå¼åè 交è°åï¼ç¬è åç°ä¸ä¸ªè¾ä¸ºæ®éçè§ç¹æ¯ï¼å½äº§å¤§æ¨¡åï¼å°¤å ¶æ¯GLM-5.2çææ°ç模åï¼å·²ç»å¯ä»¥æ¿æ¥ä¸äºè¾ å©æ§çã次è¦çä»»å¡ï¼ä¾å¦åæµè¯åä¿®æ¹bugã
ä½å¨è¾ä¸ºå¤æçãé¿æ¶é´çå¼æ¥ä»»å¡ä¸ï¼å½äº§æ¨¡åç®åä¸GPTãCaludeç顶 å°æ¨¡åï¼ä»æè¾å¤§å·®è·ã
å¼åè å°å¼ ï¼ä»å¹´3æ份ç¨è¿ä¸æ¬¡å½äº§æ¨¡åç¼ç¨ï¼ç¨GLMãqwenåGPT对æ¯åæ ·çä»»å¡ï¼åªægptè·å®äºï¼ä¸ç¬¦å页é¢è¦æ±ï¼é£æ¯ä¸ä¸ªå¤çº§ãå¤åå 容æ¯å¯¹åç»æçæç项ç®ã
èå¦ä¸ä¸ªå¼åè å°èå表示ï¼æ ¹æ®å¥¹çä½éªï¼Kimi 2.7 å è±å 2.1 Pro æ§è½ä¹å°å¯ï¼ä½æ¯å®æ»çæ¥è¯´è¿æ¯è¦å¼±ä¸æ¡£ï¼å±äºç¬¬äºæ¢¯éï¼ï¼ä½æ»ä½èè¨ä¹è¶ è¿ Claude Sonnet 4.6 äºã
ç®åï¼å¼åè 群ä½ä¸çæ®éå ±è¯æ¯ï¼GLM-5.2æ¯ç¬¬ ä¸ä¸ªè¾¾å°äº Opus 级å«çå½äº§æ¨¡åã
å®å¨æ§è¡çå®çãå¤æçãé¿æ¶é´çå¼æ¥ä»»å¡ä¸ï¼å®æ¯å¯ç¨çãå æ®äºä¸å¸ä¹å°çã
尽管å¨å¾å¤ç»´åº¦ï¼ä¾å¦ä¸çç¥è¯ä¸ï¼å ¶ä¸çæ£çOpus模åä»æè¾å¤§å·®è·ï¼ä½è¿ä¸å¦¨ç¢å ¶æ为å½äº§æ¨¡åçæ°é«åº¦ã
ä½ç°å®æ¯ï¼è¿ååå´é²å¤´è§å°æ°âå°åçâï¼å´è¢«ä¸æ§è½æ å ³çå ¶ä»å ç´ æç´¯äºã
å ·ä½æ¥è¯´ï¼è¿æ ·çå ç´ å æ¬äºç¼åæçä¸é«ãé«å³°æä»ç¶ä¼éæµççï¼è®©å¼åè çå®é ä½éªå¾ç³ç³ã
è¿å°±å¼åºäºå½ä¸æ¨¡å对æ¯ä¸çå¦ä¸ä¸ªç»´åº¦ï¼ç¨³å®æ§ã
å¨ä¸çº¿å¼åè ãç¨åºåçä½éªä¸ï¼æ¨¡åçéæ©ï¼æ©å·²ä¸æ¯ç®åçæ§è½å¯¹æ¯ï¼èæ¯ââ
æ§è½âä»·æ ¼â稳å®æ§çâé»éä¸è§âã
å¨è¿ä¸ä¸ªç»´åº¦ä¸ï¼å½äº§æ¨¡ååªå强æ¿ä¸äºç¬¬ ä¸ç»´åº¦çé¨åææ ã
è¿æï¼å½å çKimiåå¸äºå ¶ææ°çæè°æ¨¡åK3ï¼è¿æ¯ä¸æ¬¾æ¥æ2.8ä¸äº¿åæ°ãåçå¤æ¨¡æå100ä¸Tokenä¸ä¸æçå¼æ¾æéæè°æ¨¡åï¼ä¸»æ»é¿ç¨ç¼ç¨ãç¥è¯å·¥ä½ä¸æ·±åº¦æ¨çã
å¨ä¼å¤æµè¯ä¸ï¼å ¶å¨Artificial Analysisä¸çæºåææ°æ¯57åãä½åå ¨ç第ä¸ï¼Arenaå端ç¼ç¨æ¦ä»¥1679åç»é¡¶ï¼çè³æä¼ è¨ç§°ï¼å ¶æ§è½å·²ç»è½æ¯è©Claude Fable 5ã
ç¶èï¼å¨èç¼çåæ°ä¸æåèåï¼å¼åè å ³åæå¤çï¼ä»ç¶æ¯ä¸ä¸ªè¯ï¼æ§ä»·æ¯ã
å¨K3åå¸åï¼å¾å¤å¼åè 表示ï¼ç°å¨Kimiä»ç¶å¾è´µï¼ä¸Codexç¸æ¯æ²¡ææ§ä»·æ¯ï¼ä¸é¢åº¦ä¸å¤ç¨ï¼åçAPIä»·æ ¼ï¼K3ä¹è°ä¸ä¸âç½èä»·âï¼æ¯ç¾ä¸Tokenç¼å/è¾å ¥/è¾åºåå«ä¸º2/20/100å ï¼GPT-5.6 Sol为0.5/5/30ç¾å ï¼K3è½ä½äºSolï¼å´ææ¾é«äºGPT-5.6 Lunaç0.1/1/6ç¾å ã
ä¸å¨ä½¿ç¨K3çè¿ç¨ä¸ï¼å¼åè 表示è¿åºç°äºAPIæè¿ï¼çè³æäºä¸ä¸åçæ åµã
å°±ç®åçæ åµæ¥çï¼å½å å¼åè ä¸æ¯ä¸æ¿æ为å½äº§æ¨¡åä»è´¹ï¼èæ¯å½äº§æ¨¡åç®åå ç®åç´§ãææ¬é«ï¼ç®åå¼ä¸èµ·Codexè¿æ ·å ·ææ§ä»·æ¯çå¥é¤ï¼åªè½å¼çä¼¼èªå©é¤ï¼å®åééçâ大ä¼ç¹è¯å¥é¤âï¼ç±»ä¼¼æºè°± çCoding Planï¼ã
èå¥é¤çä½éªä¸ç¨³å®ãæ§ä»·æ¯æ¨¡ç³ï¼å¯¼è´ç¨æ·ä½¿ç¨è¿é¾ä»¥çåãè¿æ¬è´¨ä¸ï¼æ¯å¨ç®åç®ååºåº§ææ¬é«çæ åµä¸ï¼ç¨æ·ççæ§éæ©ã
以ä¸è¿äºï¼å¹¶éæ们è¦é¿ä»äººå¿æ°ï¼çèªå·±å¨é£ï¼åªæ¯æ³æ示ä¸ç§é£é©ã
ç®åï¼å½äº§æ¨¡åå¨è¿½èµ¶GPTãClaudeçè·¯éä¸ï¼é常å欢强è°æ§è½ï¼ä½å½äº§å¤§æ¨¡åä»ç åå°åºç¨ï¼éè¦æç»çï¼è¿ä¸æ¢æ§è½è¿ä¸åº§é«å±±ã
æ们éè¦çæ¯çå¾ ï¼çå°å½äº§ç®ååºåº§å¤§è§æ¨¡é产äºï¼å¨ææ¯ãä»·æ ¼ã稳å®æ§æ¹é¢ï¼å½¢æååäºï¼æ们ç产åå°±è½æèµ·ç¨æ·çæå¾ ã
ãæ¬æç±æèµçåä½ä¼ä¼´å¾®ä¿¡å
¬ä¼å·ï¼æè¯FunTalkææåå¸ï¼æ¬å¹³å°ä»
æä¾ä¿¡æ¯åå¨æå¡ãã
ãå
责声æãï¼æ¬æä¸ææä»»ä½æèµå»ºè®®ãå¸åºæé£é©ï¼æèµéè°¨æ
ã
å¦æä»»ä½çé®ï¼è¯·èç³»ï¼editor@zero2ipo.com.cnï¼æèµçå¤çã