ModelChorusModelChorus
KalubaleHiraJadawalin MatsayiGusarGwajiTarihi
Yadda yake aikiAPISharuɗɗan SabisManufar Sirri

Haƙƙin mallaka 2026 MeetKai Inc.

Gwaji/GPT-5 Nano/ayyuka Urdu (Pakistan)

GPT-5 Nano

ayyuka 8

Kowane jeri a ƙasa aiki ɗaya ne na tantancewa wanda aka gwada wannan samfurin a kai. Ginshiƙin Maki yana ɗaukar matsakaicin kowane ma'auni da aikin ke bayar da rahoton (daidaito, F1, daidai-daidai, da sauransu). Danna kan jeri don bincika tambayoyin daban-daban da amsoshin samfurin.

Matsakaici
59.5
MakiHarsheAikiMa'auni
89.2Urdu (Pakistan)
urdu_uquad
urdu qa
llm_judge_score: 89.2sample_len: 139.0
llm_judge_score: 89.2sample_len: 139.0
77.4Urdu (Pakistan)
urdu_bingcheck
urdu claim
f1_macro: 77.4sample_len: 102.0
f1_macro: 77.4sample_len: 102.0
75.2Urdu (Pakistan)
urdu_freshqa
urdu qa
llm_judge_score: 75.2sample_len: 323.0
llm_judge_score: 75.2sample_len: 323.0
65.3Urdu (Pakistan)
urdu_factcheckbench
urdu claim
f1_macro: 65.3sample_len: 387.0
f1_macro: 65.3sample_len: 387.0
61.6Urdu (Pakistan)
urdu_facttool_qa
urdu claim
f1_macro: 61.6sample_len: 160.0
f1_macro: 61.6sample_len: 160.0
57.6Urdu (Pakistan)
urdu_fake_news
urdu classification
f1_macro: 57.6sample_len: 300.0
f1_macro: 57.6sample_len: 300.0
28.5Urdu (Pakistan)
urdu_emotion_class
urdu classification
f1_macro: 28.5sample_len: 200.0
f1_macro: 28.5sample_len: 200.0
21.0Urdu (Pakistan)
urdu_simpleqa
urdu qa
llm_judge_score: 21.0sample_len: 200.0
llm_judge_score: 21.0sample_len: 200.0