Llama-4-Maverick-17B-128E-Instruct

MERA Создан 18.08.2026 17:03
0.226
Общий результат
27
Место в рейтинге

В топе по задачам:

10 Enantiosemy

Задачи с минимальными баллами:

31 Characters
30 GorillaHard
30 NewReasoning
29 Humor
29 SOB-Hard
Модель, команда
Результат
{{ formatValue(getLeaderboardPlace(submit)) }}
{{ submit.name }} {{ transformSize(submit.model_size) }}
{{ submit.team_name }}
{{ formatValue(getResultValue(submit)) }}
{{ column.title }} Время
{{ formatValue(getColumnValue(submit, column)) }} {{ formatValue(getTotalTime(submit)) }}

Все задачи

Задача Результат Метрика
SAGE 0.651 / 0.568 / 0.61
PUNCT_F1 SPELL_F1 ERRANT_F1
Humor 0.082 / 0.197
Exact match LLM judge score
LIMUR 0.504 / 0.589
Exact match LLM judge score
RUBIN 0.579 / 0.589
Exact match LLM judge score
Riddles 0.318 / 0.345
Exact match LLM judge score
SOB-Hard 0.019 / 0.033 / 0.72 / 0.03 / 0.036 / 0.67
balance_score task_pass_rate format_pass_rate sample_pass_rate content_pass_rate constraint_pass_rate
Characters 0.273 / 0.595
Exact match LLM judge score
Enantiosemy 0.449 / 0.547
Exact match LLM judge score
GorillaHard 0.229 / 0.3 / 0.172 / 0.254 / 0.633 / 0.783 / 0.165 / 0.205 / 0.833 / 0.008 / 0.932 / 0.859 / 0.005 / 1
balance_score clarify_recall args_match_rate tool_match_rate dialog_pass_rate format_pass_rate sample_pass_rate abstention_recall cost_optimal_rate false_clarify_rate constraint_pass_rate tool_in_catalog_rate false_abstention_rate injection_resistance_rate
IFHardBench 0.398 / 0.014 / 0.465
balance_score sample_pass_rate constraint_pass_rate
NewReasoning 0.322 / 0.396
Exact match LLM judge score
RussianRegions 0.123 / 0.437 / 0 / 0.104
Exact match LLM judge score Group Exact match Group Judge Score

Информация о сабмите

Команда:

MERA

Название ML-модели:

Llama-4-Maverick-17B-128E-Instruct

Размер модели

17.0B

Тип модели:

Открытая