UGoogle ukhuphe umboniso we Gemini 2.5 Ukusetyenziswa kweKhompyutha, imodeli eyenzelwe ukuvumela ii-arhente ze-AI ukuphatha ujongano njengomsebenzisi: jonga, cinezela amaqhosha, okanye ugcwalise iifom. Inkampani ke ngoko ivula iqonga kubaphuhlisi abaninzi, ngokusebenzisa Gemini API, ubuchwephesha obusezandleni zabavavanyi bangaphakathi ukuza kuthi ga ngoku.
Ngenxa ye Gemini 2.5 Pro, le ndlela ibeka phambili i intsebenziswano ngqo kunye nokusingqongileyo okubonakalayo (iiphequluli kwaye, kwezinye iimeko, ujongano oluphathwayo). Injongo kukuba umsebenzisi anikeze imisebenzi yemihla ngemihla ngaphandle kweprogram, ngelixa egcina ulawulo lokhuseleko kunye ne-latency ephantsi kwimisebenzi yokukhangela.
Yintoni iGemini 2.5 yokusetyenziswa kweKhompyutha kwaye yenzelwe ntoni?
Isiphakamiso sijolise ekuvumeleni i-arhente ukuba yenze izigqibo kunye yenza iintshukumo kujongano ngexesha lokwenyani. Xa kuthelekiswa nodibaniso lwe-API yemveli, usetyenziso lweKhompyutha lugubungela ezo nkqubo apho kukho i-UI yomzobo kuphela: iiphothali zokubhukisha, amaphepha aneefomu okanye iiphaneli zangaphakathi ezinobuqinisekiso, ezifanayo nezo zilawulwa Umkhangeli zincwadi weComet.
Ngokwezinto eziphathekayo, imodeli inako Ndwendwela iisayithi, jonga, chwetheza umbhalo, cofa, kwaye utsale/uwise izinto. Ikwaphatha izinto ezintsonkothileyo ezinje ngemenyu eyehlayo, izihluzi, kunye nokuhamba kokungena, okwenza kube lula ukubamba izikhokelo kwiCRM okanye iinkqubo ezipheleleyo ezisekwe kwiwebhu ukusuka ekuqaleni ukuya ekugqibeleni.
UGoogle ulungelelanisa le teknoloji kunye namanyathelo e-arhente: yinxalenye ephambili ye Iprojekthi yeMariner kwaye incedisana nenkqubela phambili ye Iprojekthi yeAstra. Ukongeza, iinguqu zemodeli sele zixhasa amandla earhente kwi Imowudi yokukhangela ye-AI, ivula indlela yamava azenzekelayo aluncedo ngakumbi kwaye anamandla.
Kuvandlakanyo lwangaphakathi nangaphandle, kubandakanywa iimvavanyo ezifana nezo ze Isiseko sebhrawuza, inkqubo ibonise ukuchaneka okuphezulu kokulawula isikhangeli ngelixa igcina Ukubambezeleka okuphantsiOku kufanelekileyo kwimisebenzi ekhonkxiwe apho isantya sichaphazela amava omsebenzisi wokugqibela.
Nangona kugxilwe kwisikhangeli sedesktop, uGoogle ujonge kuyo izigaba mobileNangona kunjalo, ulawulo lwe-desktop ye-OS alukabikho ngaphakathi kwendawo yokujonga kuqala.
Isebenza njani: amagalelo, iluphu yentshukumo, kunye nokhuseleko
Ukufikelela ngeGemini API kwiGoogle AI Studio kunye neVertex AI, usebenzisa isixhobo ukusetyenziswa_kwekhompyuthaUkuhamba kusebenzisa amagalelo amathathu angundoqo: Umyalelo womsebenzisi, umfanekiso wekhusi wokusingqongileyo, kunye nembali yezenzo zamva nje.
Ngolo lwazi, imodeli isizathu malunga nemeko yokubonwayo kwaye ibuyisela impendulo emele intshukumo ye-UI elandelayo (umzekelo, cofa okanye uchwetheze okubhaliweyo). Umxhasi uphumeza eso senzo, uhlaziya isikrini, kwaye uthumela umxholo omtsha, ukuvala i i-loop ephindaphindayo ephindwayo de umsebenzi ugqitywe.
Kwimisebenzi ebuthathaka-efana nokuthumela idatha yobuqu okanye ukuqinisekisa ukuthenga-inkqubo inakho cela isiqinisekiso esicacileyo kumsebenzisi. Ukongeza, amanyathelo acetywayo axhomekeke ku-a umvavanyi wokhuseleko lwangaphandle ehluza ukuziphatha okunokuba nobungozi okanye okungafunwayo.
Le ndlela ye-multimodal idibanisa isicatshulwa kunye nomfanekiso (i-captures) kunye nenkxaso ulandelelwano olude lwezenzoUkudityaniswa kokuqiqa okubonakalayo kunye nenkumbulo yamanyathelo angaphambili kukuvumela ukuba woyise ukuhamba komsebenzi wokwenyani kunye nokuqinisekiswa, iifom ezinemiqathango, kunye nezinto ezisebenzisanayo.
Ngokweedemos ezabiwe nguGoogle, iiarhente zinokuthi, umzekelo, lawula amadinga ukufunda idatha kwiwebhusayithi, ukucoca ngendawo, ukugcwalisa iCRM kunye nokucwangcisa i-slot; okanye tolika ibhodi yezaziso kwaye uyilungelelanise ngokutsha ngokweendidi ezichazwe kwangaphambili, ebonisa ukukwazi ukuqonda umxholo wokubonwayo.
Iimeko zokusetyenziswa kwangoku, ukufikelela kunye nemida
Kwimveliso kunye namaqela okusebenza, Ukusetyenziswa kweKhompyutha kwenza akukho-khowudi automations kwimisebenzi efana nokubhaliswa kwabathengi, ukuqinisekiswa kwedatha yeqonga, okanye ukuhlaziya uluhlu lwempahla kwiindawo ezininzi. Ekuphuhlisweni, kuluncedo kwi uvavanyo lojongano, ukuqinisekiswa kokuhamba kunye neprototyping ye-agent ekhawulezayo.
Umboniso woluntu uyafumaneka Google AI Studio (uvavanyo nophuhliso) kunye Vertex AI (ukusasazwa kwamashishini), ukongeza kwi Gemini API ukudityaniswa kwesiko. Oku kuququzelela uhlolo olulawulwayo ngaphambi kokuzisa iteknoloji kwimveliso.
UGoogle ubalaselisa iingenelo ezinje ukuncitshiswa kweendleko zokusebenza ngokuzenzekela, ukuphuculwa kokulinganisa, kunye nokukhulula ixesha lemisebenzi enexabiso eliphezulu. Kwangaxeshanye, khumbula ukuba le yinguqulelo yokujonga kwangaphambili: kunokubakho iimpazamo kunye nezithintelo kwindawo ezithile okanye amalungu e-UI.
Ngokubhekiselele kuluhlu, imodeli i ilungiselelwe umkhangeli zincwadi kwaye ucinga ngeemeko ezihambahambayo, nangona le ndawo yokugqibela inokubonisa imida ngokuxhomekeke kujongano. I ulawulo lwenkqubo yokusebenza Idesktop ayiqukwanga ngoku kwinkxaso esemthethweni.
Kwimeko yecandelo, uGoogle ukhuphisana nezisombululo ezivela kwi-real estate agents. I-OpenAI o I-Anthropic. Indlela yokuSetyenziswa kweKhompyutha, ejolise ekuqiqeni okubonakalayo, iilophu zentshukumo kunye nezikhuselo, ifuna ukubonelela ngolungelelwaniso phakathi komthamo ozenzekelayo kunye ukhuseleko lomsebenzisiKukho kwakhona izindululo zomntu wesithathu ezifana Isikhangeli se-OpenAI eziphonononga ii-arhente ezidibeneyo.
Ifoto jikelele esishiyiweyo sesibhengezo ngulo we-a oluzenzekelayo olufikelelekayo kwiinkqubo ezikhoyo kuphela njengojongano olubonakalayo, nogxininiso olucacileyo kulawulo, ukuchaneka, kunye nezikhuselo. Nabani na ofuna ukuyivavanya angayenza ngoku kwi-AI Studio okanye kwi-Vertex AI, ekhumbula ukuba oku imbonakalo kwaye ukukwazi ukuphatha indlela yokusebenza yedesktop ayikabikho kwimenyu.