MiniMax – LLM yamphamvu yotsika mtengo yopikisana ndi OpenAI pakuchita bwino komanso kupezeka mosavuta

#AI #EarnCrypto #minimax

MiniMax (yomwe yafotokozedwa pa Quasa.io/projects/minimax) ndi nsanja ya AI yotsogola yochokera ku kampani yatsopano ya ku China yotchedwa MiniMax, yomwe ikutchuka mu 2026 chifukwa cha mamodeli ake apamwamba a zinenero zazikulu ndi zida zopangira zinthu zatsopano, makamaka polemba khodi, pochita ntchito motsatizana pogwiritsa ntchito ma agent, komanso popanga mavidiyo. 

Monga mmodzi mwa otsogola popanga zinthu zatsopano za LLM ku Asia, MiniMax ikupangitsa kuti AI yapamwamba ipezeke kwa anthu ambiri potulutsa mamodeli a khodi yotseguka komanso kupereka mwayi woigwiritsa ntchito pamtengo wotsika, ndipo ikupikisana ndi makampani akuluakulu monga OpenAI ndi Anthropic pakuchita bwino ndi kuthandiza pa ntchito zenizeni.

Mphamvu yaikulu ya MiniMax ili m'mamodeli monga M2.1 — modeli yamphamvu ya khodi yotseguka yomwe yakonzedwa kuti izilemba bwino khodi m'zinenero zosiyanasiyana (kuphatikizapo Python, Java, C++, ndi zina) komanso kukhazikitsa maziko a ma agent, zomwe zimathandiza kuiphatikiza mosavuta ndi ma framework monga Claude Code, Droid, ndi Cline. Mungapange ma app omangidwa pa maziko a AI kuti agwire ntchito za kayendetsedwe ka mabungwe, sayansi ya deta, zachuma, HR, ndi kupanga mapulogalamu, mogwira ntchito bwino kuyambira pachiyambi mpaka kumapeto ngakhale pa ntchito zovuta zokhala ndi magawo ambiri. Chida chawo chopangira mavidiyo cha Hailuo 02 chimachita bwino kwambiri, ndipo chimapanga mavidiyo ooneka ngati enieni komanso ngati a m'mafilimu kuchokera ku malangizo olembedwa pamtengo wa $8/month wokha — kuposa zida zokwera mtengo monga Veo 3 pa ubwino ndi kupezeka mosavuta.

MiniMax Agent ndi chida chinanso chapadera: ndi AI yogwira ntchito zosiyanasiyana, yokwanitsa kuchita ntchito zovuta zotenga nthawi yaitali ndi malangizo ochepa chabe, ndipo imapezeka pa agent.minimax.io. Mphamvu zake zazikulu ndi kuchita bwino kwambiri kuposa ena ambiri m'gawoli pa ntchito za ma agent, kupanga mavidiyo osonyeza zinthu molondola kwambiri (kayendedwe kosalala, zithunzi zokhala ndi tsatanetsatane), luso lamphamvu logwiritsa ntchito zinenero zambiri, ndi kuika patsogolo kugwira ntchito bwino popanda kuwononga zambiri (kuyankha mofulumira, kugwiritsa ntchito zinthu zochepa za pakompyuta). Zosintha zaposachedwa za mu 2026 zawonjezera thandizo la M2.1 popanga mapulogalamu a Android/iOS, zapangitsa mavidiyo a Hailuo kuoneka ngati enieni kwambiri, komanso zawonjezera njira zabwino zophatikizira zida ndi ma agent.

Opanga mapulogalamu ndi opanga zinthu amakonda mfundo ya khodi yotseguka (“pomaliza pakhala modeli yopikisana ndi ya khodi yotsekedwa popanda mtengo wake”), kusavuta kopanga ma agent ogwirizana ndi zosowa zawo (“imasandutsa ntchito zovuta zokhala ndi magawo ambiri kukhala mayankho ongodina kamodzi”), komanso phindu la Hailuo poyerekeza ndi mtengo wake (“mavidiyo aukatswiri ndi ndalama zochepa”). Ndi yamphamvu makamaka kwa opanga mapulogalamu odziimira paokha, opanga zinthu zofalitsa, ndi mabizinesi amene akupangitsa ntchito za mu ofesi kuchitika zokha kapena kupanga zithunzi ndi mavidiyo otsatsa.

Zofooka: mamodeli ena monga M2.1 amafunika kuphunzitsidwanso mwapadera kuti azichita bwino kwambiri pa ntchito zapadera kwambiri, pali malire pa kupanga mavidiyo (mwachitsanzo, kutalika kapena kuthwa kwa chithunzi pa mapulani oyambira), ndipo zolemba za nsanjayi zopereka malangizo m'Chingelezi sizinakwanirebe (ngakhale zikuwongolera). Koma kwa aliyense amene akufufuza AI ya khodi yotseguka yolembera khodi, yogwiritsira ntchito ma agent, kapena yopangira mavidiyo, MiniMax imapereka zida zamakono kwambiri zomwe zimapezeka mosavuta kuposa zina.

Ndi nsanja yomwe ikukwera kutchuka pa nkhani za AI padziko lonse — pezani mphotho ya 1 QUA kudzera pa Quasa inunso! 

Nyenyezi 4.8/5 (ndi yapadera chifukwa cha kupanga zinthu zatsopano, kugwira ntchito bwino, ndi mitengo yotsika; yachotsedwa pang'ono chifukwa cha mipata m'zolemba zopereka malangizo komanso kufunika koiphunzitsanso mwapadera pa zochitika zina zapadera).

Yambani: https://quasa.io/projects/minimax