- I-OpenAI izokhipha i-GPT-4.5 njengemodeli yayo yokugcina yokucabanga ngaphambi kokuhlanganisa imindeni yayo ye-AI.
- I-GPT-5 izohlanganisa ubuchwepheshe obuthuthukisiwe, okuhlanganisa i-o3, futhi inikeze amazinga ahlukene okufinyelela kuye ngokubhalisa.
- Inkampani izosusa isikhethi semodeli futhi yenze lula ulwazi lomsebenzisi ngobuhlakani obuhlanganisiwe.
- Ukwethulwa kwe-GPT-4.5 ne-GPT-5 kuzokwenzeka ezinyangeni ezizayo, kodwa alukho usuku oluqinisekisiwe.
I-OpenAI inqume ukuthatha isinyathelo esibalulekile ekuthuthukisweni kwamamodeli ayo obuhlakani bokwenziwa . USam Altman, i - CEO yenkampani, umemezele imephu entsha yomgwaqo ehlose ukwenza lula nokuthuthukisa ulwazi lomsebenzisi ngokusebenzisa isu lokuhlanganisa. Kule nqubo, i-GPT-4.5 izophawula isigaba sokuguquka ngaphambi kokwethulwa kwe- GPT-5 , imodeli ezohlanganisa ubuchwepheshe obuhlukahlukene futhi ihlanganise amakhono e-AI athuthukile.
Sekuyiminyaka eminingi, i-OpenAI ithuthukise izinguqulo eziningi zamamodeli ayo olimi, okuholele ekudidekeni okuthile phakathi kwabasebenzisi bayo. Ngokufika kwe -GPT-4.5 kanye ne-GPT-5 , inkampani ihlose ukuqeda lokhu kuhlukana futhi inikeze uhlelo olusebenza kahle futhi olufinyeleleka kalula.
I-GPT-4.5: Imodeli yokugcina ngaphandle kokucabanga ngeketango
I-GPT-4.5 , eyaziwa ngaphakathi ngokuthi "i-Orion," izoba imodeli yokugcina ye-OpenAI engasebenzisi ukucabanga kweketanga . Lobu buchwepheshe buyasiza ekwakheni izimpendulo ezithuthukisiwe kakhulu emisebenzini eyinkimbinkimbi, ngakho amamodeli e-OpenAI esikhathi esizayo azohlanganisa unomphela lo msebenzi.
Le nkampani inqume ukuthi le modeli izosebenza njengebhuloho lokuhlanganisa imindeni yayo ye-AI. I-OpenAI ihlela ukuhlanganisa uchungechunge lwe-"o" kanye ne-"GPT", idale uhlelo oluguquguqukayo olusebenzisa wonke amathuluzi atholakalayo futhi olufaneleka kangcono ezimweni ezahlukene zokusetshenziswa.
I-GPT-5: Ukuhlanganiswa okugcwele nokufinyelela okunezigaba
Ukwethulwa kwe- GPT-5 kuzoletha ushintsho olukhulu endleleni abasebenzisi abafinyelela ngayo ubuhlakani bokwenziwa be-OpenAI . Esikhundleni sokugcina amamodeli ahlukene, njengoba bekulokhu kunjalo kuze kube manje, inkampani izohlanganisa ubuchwepheshe obuningi futhi ithuthukise ulwazi lomsebenzisi.
Lo modeli uzofaka ubuchwepheshe be-o3 , obekulokhu kukhuthazwa njengomkhiqizo ozimele kuze kube manje. I-OpenAI iqinisekisile ukuthi i-o3 izophela njengemodeli ehlukile futhi izoba yingxenye ye -GPT-5 . Ngaphezu kwalokho, kuzosetshenziswa uhlelo lokufinyelela olunezigaba olusekelwe ezinhlelweni ezahlukene zokubhalisa .
- Abasebenzisi bamahhala: Uzoba nokufinyelela okungenamkhawulo kukho GPT-5 in yakhe Ukumiswa okujwayelekile.
- Kanye nababhalisile: Bayoba a Uhlobo oluthuthukisiwe yemodeli, namandla amakhulu okucabanga.
- Abasebenzisi bePro: Bazoba ne izinga eliphezulu kakhulu ubuhlakani bokwenziwa, obunamandla anwetshiwe.
I-GPT-5 izophinde ifake izici ezintsha, njengamakhono ezwi , indawo ebonakalayo esebenzisanayo , amathuluzi okusesha iwebhu athuthukile, kanye nemisebenzi yocwaningo olujulile.
Ukususa isikhethi semodeli nokwenza kube lula ukuzizwisa
Enye yezinguquko eziphawuleka kakhulu i-OpenAI ezoyisebenzisa ngalesi sibuyekezo ukususwa kwesikhethi semodeli ku-ChatGPT. Le nkampani ivumile ukuthi leli thuluzi alizange lisebenze kahle ngokuphelele nokuthi abasebenzisi bavame ukuthola lokhu kudideka. Ngakho-ke, ngokukhishwa kwe- GPT-5 , ukuguquka phakathi kwamamodeli kuzophathwa ngokuzenzakalelayo ngokusekelwe emsebenzini okhona.
USam Altman uthe umgomo uwukunciphisa ubunzima kubasebenzisi nokwenza ubuhlakani bokwenziwa busebenze ngendlela eqondakala kalula nephumelelayo.
Ukuncintisana nezinselelo emakethe
Lezi zinguquko zifika ngesikhathi lapho ukuncintisana ngaphakathi komkhakha wobunhloli bokwenziwa kuya ngokuya kuba nzima. Izinkampani ezifana ne -Google , i-Microsoft , kanye nokufika kwamuva kwe- DeepSeek kufake ingcindezi ku-OpenAI, okuholele inkampani ekuthuthukiseni amasu anolaka okugcina ubuholi bayo.
Isibonelo, i-DeepSeek yethule amamodeli okucabanga ngezindleko eziphansi kakhulu, okubeka inselele ku-OpenAI. Ngenxa yalesi simo, inkampani ifuna ukusheshisa ukuthuthukiswa kwe- GPT-5 nokuthuthukisa izinsizakalo zayo ukuze ihlale ingumholi embonini.
Nakuba kungekho usuku oluthile lokukhishwa olubekiwe lwe -GPT-4.5 kanye ne-GPT-5 , kulinganiselwa ukuthi lawa mamodeli angatholakala ezinyangeni ezizayo. I-OpenAI izibophezele ekushintsheleni endaweni ye-AI efinyeleleka kalula futhi esebenza kahle , lapho ubuhlakani bokwenziwa kulula ukubusebenzisa nokubuqonda.