ModelChorusModelChorus
DéfiChatClassementBalayerBenchmarksHistorique
Comment ça marcheAPIConditions d'utilisationPolitique de confidentialité

Droits d'auteur 2026 MeetKai Inc.

Benchmarks/GPT-oss-120B/Toutes les tâches

GPT-oss-120B

67 tâches

Chaque ligne ci-dessous est une tâche d'évaluation effectuée par ce modèle. La colonne Score moyenne toutes les métriques rapportées par la tâche (précision, F1, correspondance exacte, etc.). Cliquez sur une ligne pour parcourir les questions individuelles et les réponses du modèle.

Moyenne
59.1
ScoreLangueTâcheMétriques
96.3Ukrainian (Ukraine)
ukrainian_polywrite
ukrainian open generation
open_quality_score: 96.3sample_len: 154.0
open_quality_score: 96.3sample_len: 154.0
95.0Albanian (Albania)
albanian_polywrite
albanian open generation
open_quality_score: 95.0sample_len: 155.0
open_quality_score: 95.0sample_len: 155.0
92.1Portuguese (Portugal)
portuguese_enem
portuguese mcq
exact_match: 92.1sample_len: 1432.0
exact_match: 92.1sample_len: 1432.0
91.9Ukrainian (Ukraine)
ukrainian_belebele
ukrainian mcq
f1_macro: 91.9sample_len: 900.0
f1_macro: 91.9sample_len: 900.0
91.7Albanian (Albania)
albanian_belebele
albanian mcq
f1_macro: 91.7sample_len: 900.0
f1_macro: 91.7sample_len: 900.0
90.1French (France)
french_sib200
french classification
f1_macro: 90.1sample_len: 204.0
f1_macro: 90.1sample_len: 204.0
89.0Ukrainian (Ukraine)
ukrainian_sib200
ukrainian classification
f1_macro: 89.0sample_len: 204.0
f1_macro: 89.0sample_len: 204.0
88.8Portuguese (Portugal)
portuguese_bluex
portuguese mcq
exact_match: 88.8sample_len: 724.0
exact_match: 88.8sample_len: 724.0
88.8English (US)
english_mgsm
english math
exact_match: 88.8sample_len: 250.0
exact_match: 88.8sample_len: 250.0
87.6Swahili (Tanzania)
swahili_sib200
swahili classification
f1_macro: 87.6sample_len: 204.0
f1_macro: 87.6sample_len: 204.0
87.5Albanian (Albania)
albanian_sib200
albanian classification
f1_macro: 87.5sample_len: 204.0
f1_macro: 87.5sample_len: 204.0
87.1English (US)
english_gsm8k
english math
exact_match: 87.1sample_len: 1319.0
exact_match: 87.1sample_len: 1319.0
87.1Urdu (Pakistan)
urdu_uquad
urdu qa
llm_judge_score: 87.1sample_len: 139.0
llm_judge_score: 87.1sample_len: 139.0
86.6Portuguese (Portugal)
portuguese_hatebr
portuguese classification
f1_macro: 86.6sample_len: 1400.0
f1_macro: 86.6sample_len: 1400.0
85.9Arabic (Saudi Arabia)
arabic_sib200
arabic classification
f1_macro: 85.9sample_len: 204.0
f1_macro: 85.9sample_len: 204.0
85.8Hausa (Nigeria)
hausa_sib200
hausa classification
f1_macro: 85.8sample_len: 204.0
f1_macro: 85.8sample_len: 204.0
85.4Ukrainian (Ukraine)
ukrainian_global_mmlu
ukrainian mcq
f1_macro: 85.4sample_len: 2850.0
f1_macro: 85.4sample_len: 2850.0
82.8Albanian (Albania)
albanian_global_mmlu
albanian mcq
f1_macro: 82.8sample_len: 400.0
f1_macro: 82.8sample_len: 400.0
81.8Igbo (Nigeria)
igbo_sib200
igbo classification
f1_macro: 81.8sample_len: 204.0
f1_macro: 81.8sample_len: 204.0
79.5Albanian (Albania)
albanian_aya
albanian open generation
llm_judge_score: 79.5sample_len: 200.0
llm_judge_score: 79.5sample_len: 200.0
78.6Yoruba (Nigeria)
yoruba_sib200
yoruba classification
f1_macro: 78.6sample_len: 204.0
f1_macro: 78.6sample_len: 204.0
74.4English (US)
english_mmlu_pro
english mmlu pro
exact_match: 74.4sample_len: 2100.0
exact_match: 74.4sample_len: 2100.0
73.6French (France)
french_mgsm
french math
exact_match: 73.6sample_len: 250.0
exact_match: 73.6sample_len: 250.0
69.8Portuguese (Portugal)
portuguese_tweetsentbr
portuguese classification
f1_macro: 69.8sample_len: 2010.0
f1_macro: 69.8sample_len: 2010.0
69.6Swahili (Tanzania)
swahili_afrimgsm
swahili afrimgsm
exact_match: 69.6sample_len: 250.0
exact_match: 69.6sample_len: 250.0
68.9Yoruba (Nigeria)
yoruba_naijasenti
yoruba sentiment
f1_macro: 68.9sample_len: 4515.0
f1_macro: 68.9sample_len: 4515.0
68.8Ukrainian (Ukraine)
ukrainian_zno
ukrainian mcq
f1_macro: 68.8sample_len: 751.0
f1_macro: 68.8sample_len: 751.0
67.8Hausa (Nigeria)
hausa_naijasenti
hausa sentiment
f1_macro: 67.8sample_len: 5303.0
f1_macro: 67.8sample_len: 5303.0
67.4Hausa (Nigeria)
hausa_afrixnli
hausa nli
f1_macro: 67.4sample_len: 600.0
f1_macro: 67.4sample_len: 600.0
66.9Urdu (Pakistan)
urdu_freshqa
urdu qa
llm_judge_score: 66.9sample_len: 323.0
llm_judge_score: 66.9sample_len: 323.0
65.9Portuguese (Portugal)
portuguese_hate_speech
portuguese classification
f1_macro: 65.9sample_len: 851.0
f1_macro: 65.9sample_len: 851.0
65.9Spanish (Spain)
spanish_xquad_es
spanish xquad es
exact_match: 55.3f1: 76.5sample_len: 1190.0
exact_match: 55.3f1: 76.5sample_len: 1190.0
65.2Hausa (Nigeria)
hausa_afrimgsm
hausa afrimgsm
exact_match: 65.2sample_len: 250.0
exact_match: 65.2sample_len: 250.0
65.0Swahili (Tanzania)
swahili_afrixnli
swahili nli
f1_macro: 65.0sample_len: 600.0
f1_macro: 65.0sample_len: 600.0
64.9Igbo (Nigeria)
igbo_afrixnli
igbo nli
f1_macro: 64.9sample_len: 600.0
f1_macro: 64.9sample_len: 600.0
62.7Igbo (Nigeria)
igbo_naijasenti
igbo sentiment
f1_macro: 62.7sample_len: 3682.0
f1_macro: 62.7sample_len: 3682.0
62.5French (France)
french_fquad
french qa
exact_match: 52.5f1: 72.5sample_len: 400.0
exact_match: 52.5f1: 72.5sample_len: 400.0
61.3Portuguese (Portugal)
portuguese_oab_exams
portuguese mcq
exact_match: 61.3sample_len: 2210.0
exact_match: 61.3sample_len: 2210.0
60.8Yoruba (Nigeria)
yoruba_afrimgsm
yoruba afrimgsm
exact_match: 60.8sample_len: 250.0
exact_match: 60.8sample_len: 250.0
60.5Yoruba (Nigeria)
yoruba_afrixnli
yoruba nli
f1_macro: 60.5sample_len: 600.0
f1_macro: 60.5sample_len: 600.0
58.2Ukrainian (Ukraine)
ukrainian_squad
ukrainian qa
exact_match: 46.0f1: 70.4sample_len: 3812.0
exact_match: 46.0f1: 70.4sample_len: 3812.0
53.2Igbo (Nigeria)
igbo_afrimgsm
igbo afrimgsm
exact_match: 53.2sample_len: 250.0
exact_match: 53.2sample_len: 250.0
45.9Urdu (Pakistan)
urdu_facttool_qa
urdu claim
f1_macro: 45.9sample_len: 160.0
f1_macro: 45.9sample_len: 160.0
45.9Urdu (Pakistan)
urdu_factcheckbench
urdu claim
f1_macro: 45.9sample_len: 387.0
f1_macro: 45.9sample_len: 387.0
44.3English (US)
ifeval
ifeval
inst_level_loose_acc: 56.0inst_level_strict_acc: 45.8prompt_level_loose_acc: 44.0prompt_level_strict_acc: 31.6sample_len: 541.0
inst_level_loose_acc: 56.0inst_level_strict_acc: 45.8prompt_level_loose_acc: 44.0prompt_level_strict_acc: 31.6sample_len: 541.0
44.3Urdu (Pakistan)
urdu_bingcheck
urdu claim
f1_macro: 44.3sample_len: 102.0
f1_macro: 44.3sample_len: 102.0
43.2Arabic (Saudi Arabia)
arabic_tydiqa
arabic qa
exact_match: 32.5f1: 53.9sample_len: 921.0
exact_match: 32.5f1: 53.9sample_len: 921.0
39.6English (US)
english_belebele
english mcq
f1_macro: 39.6sample_len: 900.0
f1_macro: 39.6sample_len: 900.0
37.7French (France)
french_belebele
french mcq
f1_macro: 37.7sample_len: 900.0
f1_macro: 37.7sample_len: 900.0
37.6Urdu (Pakistan)
urdu_fake_news
urdu classification
f1_macro: 37.6sample_len: 300.0
f1_macro: 37.6sample_len: 300.0
33.7Spanish (Spain)
spanish_belebele
spanish mcq
f1_macro: 33.7sample_len: 900.0
f1_macro: 33.7sample_len: 900.0
32.9Arabic (Saudi Arabia)
arabic_belebele
arabic mcq
f1_macro: 32.9sample_len: 900.0
f1_macro: 32.9sample_len: 900.0
31.2Spanish (Spain)
spanish_global_mmlu
spanish mcq
f1_macro: 31.2sample_len: 400.0
f1_macro: 31.2sample_len: 400.0
29.7French (France)
french_mmmlu
french mcq
f1_macro: 29.7sample_len: 14042.0
f1_macro: 29.7sample_len: 14042.0
29.2Igbo (Nigeria)
igbo_afriqa
igbo qa
exact_match: 24.7f1: 33.7sample_len: 409.0
exact_match: 24.7f1: 33.7sample_len: 409.0
28.6Urdu (Pakistan)
urdu_emotion_class
urdu classification
f1_macro: 28.6sample_len: 200.0
f1_macro: 28.6sample_len: 200.0
24.7Hausa (Nigeria)
hausa_afrimmlu
hausa mcq
f1_macro: 24.7sample_len: 500.0
f1_macro: 24.7sample_len: 500.0
23.5Yoruba (Nigeria)
yoruba_afrimmlu
yoruba mcq
f1_macro: 23.5sample_len: 500.0
f1_macro: 23.5sample_len: 500.0
22.8Hausa (Nigeria)
hausa_belebele
hausa mcq
f1_macro: 22.8sample_len: 900.0
f1_macro: 22.8sample_len: 900.0
22.3Igbo (Nigeria)
igbo_afrimmlu
igbo mcq
f1_macro: 22.3sample_len: 500.0
f1_macro: 22.3sample_len: 500.0
22.1Igbo (Nigeria)
igbo_belebele
igbo mcq
f1_macro: 22.1sample_len: 900.0
f1_macro: 22.1sample_len: 900.0
20.5Urdu (Pakistan)
urdu_simpleqa
urdu qa
llm_judge_score: 20.5sample_len: 200.0
llm_judge_score: 20.5sample_len: 200.0
19.1Yoruba (Nigeria)
yoruba_belebele
yoruba mcq
f1_macro: 19.1sample_len: 900.0
f1_macro: 19.1sample_len: 900.0
17.5Hausa (Nigeria)
hausa_afriqa
hausa qa
exact_match: 15.7f1: 19.4sample_len: 300.0
exact_match: 15.7f1: 19.4sample_len: 300.0
17.5Arabic (Saudi Arabia)
arabic_aratrust
arabic mcq
f1_macro: 17.5sample_len: 522.0
f1_macro: 17.5sample_len: 522.0
15.7Yoruba (Nigeria)
yoruba_afriqa
yoruba qa
exact_match: 12.7f1: 18.8sample_len: 332.0
exact_match: 12.7f1: 18.8sample_len: 332.0
14.8Arabic (Saudi Arabia)
arabic_mmlu
arabic mcq
f1_macro: 14.8sample_len: 14316.0
f1_macro: 14.8sample_len: 14316.0