- I-GPT-5.1 ihlela kabusha umndeni ngezinhlobo ezimbili ezihlukene: Esheshayo (ingxoxo esheshayo) kanye Nokucabanga (ukucabanga okujulile) nge-Auto mode ehamba ngokombuzo.
- Ukuthuthukiswa okubalulekile: ukucabanga okuguquguqukayo, ithoni yomuntu eyengeziwe, ukulandela imiyalelo, nokwenza ngendlela oyifisayo isitayela.
- Ukutholakala: izinhlelo zokuqala ezikhokhelwayo bese kuba ezamahhala; ku-API, ukwenza imephu ku-gpt-5.1-instant kanye ne-gpt-5.1-thinking.
- Ukucabanga kunikeza iwindi elibanzi (~ 196K), i-Instant ibeka phambili ukubambezeleka okuphansi; kokubili kunciphisa imfucuza yethokheni emisebenzini elula.
Ingxoxo ephathelene namamodeli e-OpenAI isiphinde yaqina ngoba, ngokusho kwesimemezelo somphakathi, i-GPT-5.1 ihlose ukuthuthukisa lokho i-GPT-5 ekwenzile kahle nokulungisa lokho okungakholisi . Asikhulumi ngokunyuka okukhulu kwamakhono, kodwa kunalokho ukubuyekezwa okugxile ekusingatheni, indlela ukucabanga okuvumelana ngayo nomsebenzi ngamunye, kanye nekhono lokwenza ngezifiso isitayela ngokulawula okukhulu kakhulu.
Uma ubulinde imodeli ehlanganisa ubuhlakani kanye nekhono, nansi: izinhlobo ezimbili ezihambisanayo (Okusheshayo kanye Nokucabanga) kanye nomzila ozenzakalelayo okunquma ukuthi uzocabanga nini kakhulu noma uqonde ephuzwini. Ngaphezu kwalokho, kunezinguquko ezisebenzayo kulabo abahlanganisa nge-API: izihlonzi zemodeli ezintsha, amafasitela omongo ahlukaniswe ngokucacile, kanye nokuthuthukiswa kokuphepha kanye nezilinganiso ezithinta ngqo amaphrojekthi omhlaba wangempela.
Iyini i-GPT-5.1, iza nini, futhi ngani manje?
I-OpenAI yethule lesi sibuyekezo ngoNovemba 12 (imithombo ehlukahlukene ikhomba u-2025), yasibeka njengokuphindaphinda ku-GPT-5 kunokuba kube umkhiqizo omusha ngokuphelele. Umgomo oshiwo ukuthuthukisa ikhwalithi yengxoxo, ukulandela imiyalelo, kanye nokucabanga okuguquguqukayo , ukuhlela kabusha umndeni ngezinhlobo ezimbili eziyinhloko kanye nokugcina imodi ye-Auto ehambisa umbuzo enjinini efanele kakhulu.
Le nkampani ikhipha izinhlelo ezikhokhelwayo ze-GPT-5.1 kuqala (Plus, Pro, Go, kanye neBhizinisi) , ngokufinyelela kwangaphambi kwesikhathi kwama-akhawunti e-Enterprise and Education kanye nokukhishwa kancane kancane kuma-akhawunti wamahhala anemikhawulo yokusetshenziswa okungenzeka. I-GPT-5 izohlala iyimodeli yefa izinyanga ezimbalwa ukuze kube lula ukuqhathaniswa nokufuduka, kanti i-GPT-5 Pro izothuthukiswa ibe yi-GPT-5.1 Pro ngokushesha nje lapho isitholakala.
Esimweni lapho ingxenye yomphakathi ibona i-GPT-5 "njengesheshayo kodwa ebanda kancane", i-GPT-5.1 izama ukuvala leso sikhala ngezimpendulo ezifudumele, ezicacile nezifanele umongo , ngaphandle kokudela ukusebenza emisebenzini eyinkimbinkimbi noma ukwandisa izindleko emisebenzini emincane.
Izinhlangothi ezimbili ze-GPT-5.1: Okusheshayo Nokucabanga (ngemodi ye-Auto)
I-OpenAI yakha uchungechunge lube izinhlobo ezimbili ezihambisanayo: i-GPT-5.1 Instant (ukusetshenziswa kwansuku zonke, ingxoxo ye-agile, umhlahlandlela wemiyalelo engcono) kanye ne -GPT-5.1 Thinking (ukucabangela okujulile, izincazelo ezicacile kanye nolimi oluncane) . I-router, evame ukubizwa ngokuthi i-Auto, isebenza kuzo zombili, ikwazi ukukhetha ngokuguquguqukayo ukuthi iyiphi injini ezosetshenziswa kuye ngombuzo.
I-Instant iklanyelwe ukuxoxa ngokwemvelo nokuphendula ngokushesha lapho isicelo sakho singadingi ukucatshangelwa okuningi. Isici sayo esiyinhloko uhlelo lokucabanga oluguquguqukayo "olulula" olunquma ukuthi kufanele ucabange nini kakhudlwana ngaphambi kokuphendula, ngenkathi ugwema ukucubungula imisebenzi elula ngokweqile.
Ukucabanga, ngakolunye uhlangothi, kugcizelela ukucabanga: kulungisa ngokunembile isikhathi esichithwa ezinqubweni zokucabanga zangaphakathi ngokusekelwe ebunzimeni benkinga. Umphumela uba izimpendulo ezijulile uma kudingeka kanye nezimpendulo ezisheshayo lapho inselele ilula, kusetshenziswa ulimi olungaqondakali kangako kunasezinguqulweni zangaphambilini.
Imodi ezenzakalelayo isebenzisa izinkomba zomlando wengxoxo ezisheshayo , kanye namaphethini afundiwe mayelana nokuthi iyiphi imodeli exazulula kangcono izinkinga ezifanayo, ukuze inqume ukuthi "icabange kakhulu" noma iphendule ngokushesha.
Ukuqhathanisa okuqondile: izinhloso, isivinini, isitayela nomongo
Ukuze ubone ngokushesha umehluko, kuyasiza ukubheka izigaba ezisebenzayo: inhloso, ukuziphatha kokucabanga, ukubambezeleka, isitayela sempendulo, kanye nefasitela lomongo . Ekusetshenzisweni kwansuku zonke, lezi zici zinquma ukuthi yikuphi ukuhluka okusebenza kangcono kuwe.
| Isigaba | I-GPT-5.1 Instant | GPT-5.1 Ukucabanga |
|---|---|---|
| Injongo | Ingxoxo esheshayoUkulandela okuthembekile kwemiyalelo nemisebenzi yansuku zonke | Ukuhlaziywa kwezigaba eziningiizinkinga eziyinkimbinkimbi nokucabanga okujulile |
| Ukubonisana | Ukukhanya okuguquguqukayoNquma ukuthi uzocabanga nini kakhudlwana | Ukuxoxisana okuqondileIsikhathi sokucabanga silingana nobunzima |
| Isivinini | Ukubambezeleka okuphansi kakhulu njengokubalulekile | eliguqukayo: iyashesha ngezinto ezilula, ihamba kancane ngezinto eziyinkimbinkimbi |
| Esilo | Okuqondile kanye nobunganeelungiselelwe ingxoxo | Izincazelo ezihlelekileI-jargon encane nokucaca okukhulu |
| Umqondo | Amawindi aminyene kakhulu kuya ngohlelo (isb. 16K/32K/128K) | Umongo obanzi kufika ku-~196K amathokheni |
| Kungcono | Imibono esheshayoukubhala kafushane, isifinyezo esifushane, ikhodi encane | InvestigaciónUkuhlolwa kwekhodi, ukuhlaziya imibhalo ebanzi |
| Okuzenzakalelayo | Inketho yokuzenzakalelayo ekubonisaneni okuningi | Iyasebenza emisebenzini eyinkimbinkimbi ngokucacile |
| Ukukhetha ngesandla | Kuyafaneleka ngesivinini esiphezulu | Kuyafaneleka; kwezinye izinhlelo, nezabelo zamasonto onke |
| Ukunemba/Ukujula | Altakodwa ibeka phambili isivinini | Ubukhulu ngezinkinga ezinde noma ezisontekile |
| Isinxephezelo | ⚡ Isivinini > Ukujula | 🧠 Ukujula > Isivinini |
Yini eshintsha ngempela uma iqhathaniswa ne-GPT-5 (nohambo olusuka ku-GPT-4)
Ivektha yokuqala ukucabanga okuguquguqukayo . Uma kuqhathaniswa nokucabanga okunwetshiwe kwe-GPT-5 (ikakhulukazi kumamodi okucabanga kanye ne-Pro), i-GPT-5.1 inquma ngokunembile ukuthi kufanele icabange okungakanani esimweni ngasinye : okuncane ezintweni ezincane, okuningi kweziyinkimbinkimbi. Lokhu kuphumela kumathokheni angachithwanga kakhulu kanye nezikhathi zokuphendula ezihambisana nobunzima.
Okwesibili, isitayela sokuxoxa . Okuhlangenwe nakho okuzenzakalelayo manje sekufudumele futhi kungokobuntu; futhi uma usebenzisa Ukucabanga, ulimi olusetshenziswayo luyancishiswa futhi izincazelo ziyacaca ngaphandle kokulahla ubukhali. Kulabo abasebenzisa imodeli nsuku zonke, lokhu kushintsha kwezwi kuqeda ukungqubuzana.
Okwesithathu, ukwenza kube ngokwakho . I-GPT-5.1 ifaka uhlelo oluyinkimbinkimbi kakhulu lokusetha ubuntu bomsizi: ungakhetha amathoni achazwe kusengaphambili noma ulungise izici ezifana nokufingqa noma izinga lokusondelana, futhi ulawule ngisho nemininingwane ethakazelisayo njengemvamisa yezithonjana.
Ukusebenza, amabhentshimakhi, nezindleko zomhlaba wangempela zokusebenzisa
I-OpenAI ibika ukwenyuka kwezivivinyo ezifana ne -AIME 2025 kanye nezinselele zokuhlela zohlobo lwe-Codeforces , ukugcina noma ukuthuthukisa ukusebenza kwe-GPT-5 emisebenzini eyinkimbinkimbi ngokusetshenziswa kwethokheni okuphumelelayo ngenxa yokucabanga okuguquguqukayo. Ngaphansi kwemithwalo yemisebenzi exubile, Ukucabanga kungaba ngokushesha okuphindwe kabili kunokucabanga kwe-GPT-5 ezimweni ezilula futhi kuthathe isikhathi esiningi lapho inkinga ikufuna.
Ngale kwerekhodi elithile, isihluthulelo ukuthi libonakala kanjani ebhilini lakho kanye nezikhathi zakho zokucubungula. Amandla okucubungula angachithwanga kakhulu emibuzweni elula asho amathokheni ambalwa kanye nokubambezeleka okungadingekile okuncane. Kumapayipi anezicelo eziningi zansuku zonke, lokhu kulungiswa kuholela ekuzinzeni okubonakalayo kanye nezindleko ezingabikezelwa.
Ezindaweni zobungcweti, kuyabonakala intuthuko ekubhaleni amakhodi, izibalo, kanye nokucabanga ngesinyathelo ngesinyathelo , kanye nokwehla kolimi lobuchwepheshe ekucabangeni okusiza ukuqonda ngamaphrofayili angewona ochwepheshe.
Izilawuli zethoni nezobuntu: izinketho nokulungiswa kahle
I-OpenAI ingeza isikhethi sesitayela esinezinhlobo ezifana ne -Default, Friendly, Efficient, Professional, Sincere, kanye ne-Original , futhi igcina amaphrofayili afana ne-Nerd kanye ne-Cynical. Ngaphezu kwalokho, ezinye izixhumi zibonisa amalebula afana ne -Simple/Direct noma i-Enthusiastic ene-twist ehlukile , futhi ikuvumela ukuthi ulungise ukuthi izimpendulo kufanele zibe zimfushane noma zomuntu siqu kangakanani.
Lolu ngqimba lokwenza ngokwezifiso alushintshi amakhono emodeli, kodwa luvumelanisa kangcono izwi lomsizi nesimo ngasinye sokusetshenziswa : kusukela kusevisi yamakhasimende esemthethweni kuya kokuqukethwe okuthakazelisayo kakhulu. Kumabhrendi, amaqembu okusekela, noma ukuthengisa, kuyisinyathelo esiqhubekayo ekuhambisaneni nasekulawuleni.
- Amaphrofayili ajwayelekile: Ngokuzenzakalelayo (ibhalansi), Inobungane (ifudumele futhi iyakhuluma), Isebenza kahle (imfishane futhi iqondile), Ichwepheshile (ehlelekile futhi inembe), Iqotho (evulekile), Yoqobo (yokudala).
- Amanye amalebula abonakalayo: Kulula futhi kuqondile; Inomdlandla ngolunye unqenqema; Ukulondolozwa kwe-Nerd kanye neCynical.
Amawindi omongo nokugcinwa ezingxoxweni ezinde
Ukuphathwa kokuqukethwe nakho kuthuthukisiwe. I-GPT-5 isivele inwetshiwe ku-GPT-4 , futhi i-GPT-5.1 izuza leso sisekelo ngokulungiswa kokuziphatha: I-Instant ngokuvamile inikeza amafasitela amancane kuye ngohlelo (isb., i-16K ku-Free, i-32K ku-Plus/Business, kanye ne-128K ku-Pro/Enterprise) , kuyilapho i-Thinking ihlose amafasitela amakhulu aseduze namathokheni angu-196K ukuze kuhlaziywe kabanzi.
Ngaphezu komthamo ongahluziwe, ukugcinwa komongo emicu emide kuzinzile kakhulu , kunciphisa ukuhlukana kokuhambisana ezingxoxweni eziphendula izikhathi eziningi. Lokhu kuwusizo kakhulu ekusekeleni , ezisekelweni zolwazi, kanye nasezinqubweni zangaphakathi zezigaba eziningi.
Ukuphepha, ukuhlolwa kokukhiqiza, nezinguquko zokuziphatha
I-OpenAI ikhombisa ukuthuthuka noma ukulingana kumamethrikhi okuphepha ezigabeni ezifana nokuhlukumeza, inkulumo enenzondo, kanye nokufakwa kwesithombe ku-Instant variant, ngekhadi lesistimu elifaka amathebula okuqhathanisa nokuphindaphinda kwangaphambilini. Ekucabangeni, ukuphepha kufana namamodeli wangaphambilini , nokubuyela emuva okuncane ezigabeni ezithile eziqashwe.
Ukuhlanganiswa kokufudumala okukhulu nokulawulwa kobuntu okwengeziwe kudinga imingcele eqinisiwe: ukuhlolwa kwempilo yengqondo kanye nokuthembela ngokomzwelo kuyandiswa , futhi kugcinwa izivikelo ngokumelene nebhayoloji eyingozi, ukukhathazeka ngokuphepha, kanye nokwaziswa okungaqondile. Ngamafuphi, ukusunduzela "kumuntu" kuza nezivikelo ezengeziwe.
Ukutholakala ku-ChatGPT naku-API: amamodeli, ama-ID, noshintsho
Ku-interface ye-ChatGPT, abasebenzisi abakhokhelwayo bazobona i-ChatGPT 5.1 isebenza ngokukhetha phakathi kwe -Instant , Thinking , noma i-Auto mode . Ukukhishwa kuzofika kuma-akhawunti wamahhala kamuva , mhlawumbe kunemikhawulo. Ushintsho luzogcina i-GPT-5 njengohlelo lwesikhathi esidlule cishe izinyanga ezintathu.
Ku-API, imephu yokuqala ekhonjiswe yi-OpenAI ihlanganisa i-gpt-5.1-chat-latest → gpt-5.1-instant kanye ne -gpt-5.1 → gpt-5.1-thinking , okuveza ukucabanga okuguquguqukayo ezindaweni zokugcina zengxoxo. i-gpt-5.1-instant ivelele ngokuqina kwayo ekukhiqizeni, kanye ne -gpt-5.1-thinking ngokucabanga kwayo okulungisiwe kahle.
I-OpenAI iphinde yabonisa ukuthi i-GPT-5 Pro izobuyekezwa ibe yi-GPT-5.1 Pro maduze. Okwamanje, amaqembu angaqhubeka nokuqhathanisa ukusebenza kwayo namamodeli angaphambilini kumenyu ethi "Amamodeli Efa".
Umthelela osebenzayo ngephrofayili: okuqukethwe, ukumaketha, ukuhlela, nokuhlaziya
Kulabo abasebenza ngombhalo (ababhali bemibhalo, ababhali bemibhalo, abahleli), i-Instant ilula futhi ihambisana nefomethi , kuyilapho i-Thinking ingcono ekuhlukaniseni ukuhlaziywa okude noma izimpikiswano eziyinkimbinkimbi. Ukulawulwa kobuntu obusha kuletha ithoni yomsizi eduze kwezwi lomkhiqizo ngaphandle kokulahla ukunemba.
Ekuhleleni izinhlelo, Ukucabanga kukhanya ekulungiseni amaphutha, ukubuyekeza izindawo zokugcina ezinezimo ezinde, kanye nokuchaza izinqumo ngolimi oluncane; Okusheshayo kusheshisa imisebenzi emifushane nephindaphindayo. Kokuhlaziya kanye nebhizinisi, ukucabanga okuguquguqukayo kunikeza izimpendulo eziqinile ezimweni eziningi , kugxile emzamweni lapho kwenza khona umehluko ngempela.
I-FAQ Esheshayo
Yiziphi izici ezintsha eziyinhloko ze-GPT-5.1?
Izinhlobo ezimbili (Okusheshayo Nokucabanga), ukucabanga okuguquguqukayo , ithoni yomuntu , kanye nokwenza ngokwezifiso isitayela esiyinkimbinkimbi.
Kwehluka kanjani Ukusheshayo Nokucabanga?
Okusheshayo kubeka phambili isivinini kanye nengxoxo ngokucabanga okulula; Ukucabanga kucabangela kakhulu ngokwendlela eyinkimbinkimbi, ngezincazelo ezicacile kanye nolimi oluncane.
Ingabe isiyatholakala kuwo wonke umuntu?
Kuqala kudluliselwe ezinhlelweni ezikhokhelwayo bese kufinyelela kuma-akhawunti wamahhala anemikhawulo.
I-Ecosystem, izinkampani zangaphandle nokunye ukufinyelela
Ngaphandle kweziteshi ezisemthethweni, abanye abahlinzeki banikeza ukufinyelela okunye noma amanani. Amapulatifomu afana ne-CometAPI athi anikeza amamodeli akamuva ngentengo ephansi kunesemthethweni futhi ancoma ukungena ngemvume nokukhiqiza ukhiye wakho ngaphambi kokuhlanganisa. Njengokuhlala kunjalo, qinisekisa ukutholakala kwangempela kanye nemigomo yokusebenzisa ngaphambi kokusekela umkhiqizo wakho kumuntu wesithathu.
Uzobona futhi izindatshana nemiphakathi ku-X, Discord, noma i-VK ukwabelana ngokuqhathaniswa kanye nemiyalelo. Sebenzisa lokhu ukuze ukale okulindelwe , kodwa khumbula ukuthi indawo ngayinye inezici zayo (idatha, amathuluzi, imikhawulo yomongo) ezingashintsha imiphumela.
Ama-Startups nabasunguli: izinsuku zokugcina, ukusebenza kahle kanye namathuba
Imibiko yangaphambi kokumenyezelwa ikhulume ngezinsuku zokukhishwa ezilinganisiwe ngasekupheleni kukaNovemba kanye nentuthuko ekubambezelekeni nasekuphathweni komongo. Njengoba ukuqaliswa kuqhubeka, okubaluleke kakhulu kunkampani entsha ukusebenza kahle okusebenzayo : izindleko eziphansi ngomsebenzi olula, ukujula okukhulu kwemininingwane lapho kubaluleke khona, kanye nokunakekelwa okuncane kwemodeli ngenxa yokulawula isitayela kanye nokulandelwa kwemiyalelo okuthuthukisiwe.
Ku-SaaS kanye nokusebenza kwangaphakathi, lokhu kwenza kube nokwenzeka okuhlangenwe nakho komsebenzisi okujabulisayo , ama-chatbot anamathela njalo kumafomethi, kanye nama-ejenti angacabangi kakhulu ngezinto ngokungadingekile. Uma uthengisa eLatin America, ukuguquguquka kwezilimi eziningi kanye nethoni yemvelo kuthuthukisa ukwamukelwa.
Ingena kanjani kusitaki sakho: ukukhetha kwemodeli nokugeleza
Uma ungafuni ukwenza izinto zibe nzima, vele ushiye kumodi ye-Auto . Ngemithwalo yemisebenzi echazwe kahle, cindezela i-Instant ekusebenzeni okuyinkimbinkimbi okuphansi bese usebenzisa i-Thinking ngezinyathelo zokucabanga ezibucayi (isb., ukuhlolwa kwe-hypothesis noma ukuhlolwa). Ku-API, qapha ukusetshenziswa kwemali bese ulungisa imikhawulo yethokheni ngokuya ngohlobo lomsebenzi.
Ezinhlanganweni ezidinga ukubonisana nokuthuthukiswa ngokwezifiso, kunezihlanganisi ezikhethekile. Izinkampani ezifana ne-Q2BSTUDIO zinikeza izinsizakalo ezifana nama-ejenti e-AI, isofthiwe ngokwezifiso, i-BI ene-Power BI, ukuphepha kwe-cyber/ukuhlolwa kwe-pentesting, kanye nokufakwa kwamafu (AWS/Azure) okuhloswe ngakho ukuletha amamodeli afana ne-GPT-5.1 ekukhiqizweni ngokuphephile nangokukhula.
Imininingwane yobuchwepheshe nemikhuba emihle okufanele ikhunjulwe
Ezicelweni zakho, chaza ngokucacile umgomo kanye nemingcele , bese uvumela imodeli ukuthi ivumelane nokucabanga kwayo. Gwema imiyalelo eminingi ngokweqile: I-GPT-5.1 ingcono ekulandeleni amafomethi kanye nemingcele (amagama, isakhiwo, izitayela), okunciphisa ukuphindaphinda okungadingekile.
Emisebenzini yokusebenza enezigaba eziningi, hlanganisa izifinyezo ezingaphelele nezinkomba zemicu ukuze uphathe kahle umongo. Uma ukusetshenziswa kwakho kuncike kumongo omkhulu, i-Thinking with wide window izokunikeza indawo eyengeziwe; emigqeni yemvamisa ephezulu, i-Instant izokunikeza ukubambezeleka okudingayo.
Yiziphi izivivinyo nomphakathi ezikushoyo mayelana nokujula kokucabanga
Kumathematika okuncintisanayo kanye nokubhala amakhodi, kucashunwe intuthuko ngaphezu kwe-GPT-5 (isb., izinselele ze-AIME 2025 kanye nohlobo lwe-Codeforces ). Ekucabangeni okungekona okwezibalo, akukabikho ukuvumelana okuqondile , futhi abanye abasebenzisi abathuthukile bayaqhubeka nokwenza izivivinyo ze-A/B phakathi kokucabanga kwe-GPT-5.1 kanye nezinhlobo ze-GPT-5 Pro ukuze baqhathanise izici zokuhlaziywa okungaqondakali.
Umbono ojwayelekile uwukuthi i-GPT-5.1 "icabanga" kahle kakhulu uma kudingeka futhi ayichithi isikhathi uma ingadingeki. Kodwa-ke, njenganoma iyiphi i-LLM, isengenza amaphutha , futhi kuyalulekwa ukuqinisekisa izimpendulo ezizindeni ezibucayi.
Amamodeli, ama-ID, namanothi okusebenzisa
Gcina izihlonzi ezilandelayo ziseduze: i-gpt-5.1-instant (isipiliyoni sengxoxo esizenzakalelayo), i-gpt-5.1-thinking (deep reasoning), kanye ne-API mappings gpt-5.1-chat-latest → Instant kanye ne -gpt-5.1 → Thinking . Ngokushintsha, i-GPT-5 izotholakala njengohlelo lwefa ngenkathi uqhathanisa ukuziphatha futhi uhlela ukufuduka kwakho.
Kumapulani wamahhala noma aphakathi nendawo, lindela amafasitela omongo alinganiselwe kakhulu kanye nemikhawulo yokusetshenziswa kwamasonto onke kwe-Thinking. Kumasethingi ebhizinisi, sebenzisa izinketho zokwenza ngokwezifiso ukuze uvumelanise ithoni nomkhiqizo kanye nezitayela zamadokhumenti namathempulethi ukuze yonke inhlangano ikhiqize umphumela ohambisanayo.
Okokugcina, kubalulekile ukuqaphela ukuthi i-OpenAI iqinisa amakhadi ayo esistimu kanye nezilinganiso zokuphepha ngokuphindaphinda ngakunye, yize ingashicileli imininingwane ephelele yokwakha noma idatha yokuqeqesha. Iphatha imodeli njengomsizi onamandla osebenzisana nawe , hhayi njenge-oracle engenaphutha.
Noma ubani oye wabhekana nezimpendulo "eziqondile" ku-GPT-5 uzobona ngokushesha ukuthi i-GPT-5.1 ithola imvelo nokulawula ngaphandle kokulahlekelwa amandla . Phakathi kwe-Instant yemisebenzi yansuku zonke, i-Thinking for hard count, kanye ne-Auto enquma ukuthi izosheshisa nini, uhlelo lunikeza ibhalansi ebonakala kokubili engxoxweni kanye nasekubalweni kwamathokheni.
