ModelChorusModelChorus
چیلنجچیٹلیڈربورڈسوائپ کریںبینچ مارکستاریخ
یہ کیسے کام کرتا ہےAPIسروس کی شرائطرازداری کی پالیسی

کاپی رائٹ 2026 MeetKai Inc.

بینچ مارکس/Functionary Swahili Large/تمام ٹاسک

Functionary Swahili Large

29 ٹاسک

نیچے ہر قطار ایک واحد بینچ مارک ٹاسک ہے جس پر اس ماڈل کا جائزہ لیا گیا۔ اسکور کالم اس ٹاسک کی رپورٹ کردہ ہر میٹرک (درستگی، F1، exact-match وغیرہ) کی اوسط نکالتا ہے۔ انفرادی سوالات اور ماڈل کے جوابات دیکھنے کے لیے کسی قطار پر کلک کریں۔

اوسط
67.7
اسکورزبانٹاسکمیٹرکس
92.2Albanian (Albania)
albanian_belebele
albanian mcq
f1_macro: 92.2sample_len: 900.0
f1_macro: 92.2sample_len: 900.0
91.7Ukrainian (Ukraine)
ukrainian_sib200
ukrainian classification
f1_macro: 91.7sample_len: 204.0
f1_macro: 91.7sample_len: 204.0
90.9French (France)
french_sib200
french classification
f1_macro: 90.9sample_len: 204.0
f1_macro: 90.9sample_len: 204.0
89.9Swahili (Tanzania)
swahili_sib200
swahili classification
f1_macro: 89.9sample_len: 204.0
f1_macro: 89.9sample_len: 204.0
87.8Arabic (Saudi Arabia)
arabic_sib200
arabic classification
f1_macro: 87.8sample_len: 204.0
f1_macro: 87.8sample_len: 204.0
85.2Hausa (Nigeria)
hausa_sib200
hausa classification
f1_macro: 85.2sample_len: 204.0
f1_macro: 85.2sample_len: 204.0
84.8English (US)
english_mgsm
english math
exact_match: 84.8sample_len: 250.0
exact_match: 84.8sample_len: 250.0
83.2Albanian (Albania)
albanian_global_mmlu
albanian mcq
f1_macro: 83.2sample_len: 400.0
f1_macro: 83.2sample_len: 400.0
82.6English (US)
english_gsm8k
english math
exact_match: 82.6sample_len: 1319.0
exact_match: 82.6sample_len: 1319.0
82.1Albanian (Albania)
albanian_sib200
albanian classification
f1_macro: 82.1sample_len: 204.0
f1_macro: 82.1sample_len: 204.0
79.1Urdu (Pakistan)
urdu_uquad
urdu qa
llm_judge_score: 79.1sample_len: 139.0
llm_judge_score: 79.1sample_len: 139.0
74.7Portuguese (Portugal)
portuguese_hatebr
portuguese classification
f1_macro: 74.7sample_len: 1400.0
f1_macro: 74.7sample_len: 1400.0
73.7Portuguese (Portugal)
portuguese_tweetsentbr
portuguese classification
f1_macro: 73.7sample_len: 2010.0
f1_macro: 73.7sample_len: 2010.0
72.4Urdu (Pakistan)
urdu_factcheckbench
urdu claim
f1_macro: 72.4sample_len: 387.0
f1_macro: 72.4sample_len: 387.0
72.0French (France)
french_mgsm
french math
exact_match: 72.0sample_len: 250.0
exact_match: 72.0sample_len: 250.0
71.4Urdu (Pakistan)
urdu_fake_news
urdu classification
f1_macro: 71.4sample_len: 300.0
f1_macro: 71.4sample_len: 300.0
70.7Urdu (Pakistan)
urdu_bingcheck
urdu claim
f1_macro: 70.7sample_len: 102.0
f1_macro: 70.7sample_len: 102.0
69.0Urdu (Pakistan)
urdu_facttool_qa
urdu claim
f1_macro: 69.0sample_len: 160.0
f1_macro: 69.0sample_len: 160.0
67.5French (France)
french_fquad
french qa
exact_match: 57.0f1: 77.9sample_len: 400.0
exact_match: 57.0f1: 77.9sample_len: 400.0
62.8Spanish (Spain)
spanish_xquad_es
spanish xquad es
exact_match: 52.7f1: 72.9sample_len: 1190.0
exact_match: 52.7f1: 72.9sample_len: 1190.0
59.6Portuguese (Portugal)
portuguese_hate_speech
portuguese classification
f1_macro: 59.6sample_len: 851.0
f1_macro: 59.6sample_len: 851.0
58.6Ukrainian (Ukraine)
ukrainian_squad
ukrainian qa
exact_match: 47.0f1: 70.3sample_len: 3812.0
exact_match: 47.0f1: 70.3sample_len: 3812.0
56.7Urdu (Pakistan)
urdu_freshqa
urdu qa
llm_judge_score: 56.7sample_len: 323.0
llm_judge_score: 56.7sample_len: 323.0
51.8Hausa (Nigeria)
hausa_afrixnli
hausa nli
f1_macro: 51.8sample_len: 600.0
f1_macro: 51.8sample_len: 600.0
51.1Arabic (Saudi Arabia)
arabic_tydiqa
arabic qa
exact_match: 38.3f1: 63.9sample_len: 921.0
exact_match: 38.3f1: 63.9sample_len: 921.0
38.0Hausa (Nigeria)
hausa_afrimgsm
hausa afrimgsm
exact_match: 38.0sample_len: 250.0
exact_match: 38.0sample_len: 250.0
31.5Urdu (Pakistan)
urdu_emotion_class
urdu classification
f1_macro: 31.5sample_len: 200.0
f1_macro: 31.5sample_len: 200.0
22.0Urdu (Pakistan)
urdu_simpleqa
urdu qa
llm_judge_score: 22.0sample_len: 200.0
llm_judge_score: 22.0sample_len: 200.0
9.3Hausa (Nigeria)
hausa_afriqa
hausa qa
exact_match: 7.7f1: 10.8sample_len: 300.0
exact_match: 7.7f1: 10.8sample_len: 300.0