2016å¹´ã«ã¹ãã 対çã®ä¸ç°ã¨ãã¦ãã¯ã¦ã©ãäººéæ§ã»ã³ã¿ã¼ããå°å ¥ãããã¾ããããç¾å¨ã§ã¯ã¹ãã 対çã¯å¥ã®æ¹æ³ã«ããå®ç¾ããã¦ãããå©ç¨ã®å®æ ããªããã¨ããããã®ãã³åã»ã³ã¿ã¼ãçµäºãããã¨ã¨ãããã¾ããã
é·ããã®ãæé¡§ãããã¨ããããã¾ããããããããã¯ã¦ã©ãããããããé¡ããããã¾ãã
ãã¾ã: LLMã«è§£ããã¦ã¿ãã¨
ãã£ãããªã®ã§å ¨åé¡ã主è¦ãªLLMã«åçããã¦ã¿ã¾ããã
対象㯠gpt-4oãgemini-2.0-flashãclaude-3.7-sonnet ã®3種ã
è¨å®ã¯ä»¥ä¸ã®ãããªæãã§ãã
- ã·ã¹ãã ããã³ãã:
ããªãã¯ãªããªãã«çããã¢ã·ã¹ã¿ã³ãã§ãããªããªããä¸ããããããç°¡æ½ã«åçã®ã¿ãè¿ãã¦ãã ãããåçã¯æ¥æ¬èªã§è¡ã£ã¦ãã ãããæ¬å¼§ãå¥èªç¹ã¯ä¸è¦ã - temperature: 0
- æå¤§ãã¼ã¯ã³æ°: 50
çµæã¨ãã¦ã¯ã以ä¸ã®ãããªæ£ççã¨ãªãã¾ããã
- gpt-4o: 79%
- gemini-2.0-flash: 69%
- claude-3.7-sonnet: 62%
è峿·±ãã£ãåçãããã¤ãç´¹ä»ãã¾ãã
Q: æ ªå¼ä¼ç¤¾ã¯ã¦ãªã¯2021å¹´ã§åµç«ä½å¨å¹´ï¼
- gpt-4o: 20å¨å¹´ âï¸
- gemini-2.0-flash: 20å¨å¹´ âï¸
- claude-3.7-sonnet: 20å¨å¹´ âï¸
ã»ããåç´ãªç¥èåé¡ã«ã¤ãã¦ã¯å¼·ãã£ãã§ãã
Q: ã¹ã©ãã·ã¥ã¡ã¿ã«å天çããã¹ããå«ã¾ããªããã³ãã¯ï¼ï¼æ£è§£: ã¡ã¿ãªã«ï¼
- gpt-4o: ã¡ã¬ãã¹ âï¸
- gemini-2.0-flash: ã¢ã³ã¹ã©ãã¯ã¹ âï¸
- claude-3.7-sonnet: ã¡ã¬ãã¹ âï¸
ç¥èã¯ãããã®ã®ãæååä½ã®å¤å®ã¯é£ããã£ãããã§ãã
Q: ãç¾å°å¥³æ¦å£«âã¼âã¼âã¼âããâã«å ¥ãæåã¯ï¼ï¼æ£è§£: ã»ã©ã ã³ãã»ã¼ã©ã¼ã ã¼ã³ï¼
- gpt-4o: ã»ã¼ã©ã¼ã ã¼ã³ âï¸
- gemini-2.0-flash: ã»ã¼ã©ã¼ã ã¼ã³ âï¸
- claude-3.7-sonnet: ã»ã©ã ã³ âï¸
䏿¹ã§ãã¡ãã¯å ¨å¡æ£è§£ãç¾å°å¥³æ¦å£«ã®ãã³ãã®ããããï¼
Q: ãã¬ãã®ãããããã£ã¦ä½ï¼ï¼æ£è§£: ããï¼
- gpt-4o: ãã¬ã âï¸
- gemini-2.0-flash: ãã³ãã³ âï¸
- claude-3.7-sonnet: ãã¬ãã®ããªã âï¸
ããããã§ããã
Q: ã¯ã©ã ãã³ã¯ã©ãããã£ãï¼ï¼æ£è§£: ãã·ãã·ï¼
- gpt-4o: ã¯ã©ã ãã³ã¯ãã·ãã·ããã£ã âï¸
- gemini-2.0-flash: ã±ã£ã±ã£ã± âï¸
- claude-3.7-sonnet: ã´ããã´ãã âï¸
ããããã§ããã