이미지 텍스트 확인
Comparative evaluation Of Gemini and OpenAl Deep Research
Percentage Of queries Where human testers prelerred Gemini Deep Research (powered by 25 Pro Experimental) compared to OpenAl Deep Research
69.9%
Deup Rescnrch wlin Geminl 25Pro
Overall
Eperimgnital
30.196
Openall Deep Research
Instruction
following
39.49
76.996
Comprehensiveness
23.1%6
73.39
Completeness
26.796
58.2%
Writing quallty
41.89
Eriluniicnsot camtpnarescaichGurrirrepannn praiassicnalle] mEaiarmicresoreh IndusiryrscDrchlomperonnliegtrasl snapp ralucecasraiaubns Gnhinlcnslntructontolowin]
자어 대 기 다 어 기
nJeliiy
Ue spealledmstrustonsund constrams Conpenensiveness meosurs beadlhanorarge
mmorwaroncovetedinteponse adorcssilgtie Stope oTu3erte9uE-
‘Commprtanorimmmeartre-ith dlcpiandlharctahne s Oliniorimnitioniontohicoordressra
tnerenar Writioqua ty ealuaterclirity cGnGSnraaloncal orQmlizat onona Gvernlreodahilityol
대SP-
