Skip to content

MMMLU

A multilingual MMLU-style benchmark reported in provider evaluation tables.

11 models

Higher is better.

  1. 1Interfaze BetaInterfaze90.9
  2. 2Qwen3.7 MaxAlibaba90.3
  3. 3Qwen3.7 PlusAlibaba89.0
  4. 4Claude 4 OpusAnthropic87.4
  5. 5K-EXAONE 2.0LG AI Research86.6
  6. 6Gemma 4 12BGoogle83.4
  7. 7Claude 3.7 SonnetAnthropic83.2
  8. 8QwQ-32BAlibaba80.9
  9. 9Qwen3 235B-A22BAlibaba79.8
  10. 10Qwen3 32BAlibaba76.5
  11. 11Phi-3.5-miniMicrosoft55.4
Higher is better. Bar colour: blue proprietary, orange open weights, green other.

About this benchmark

Metric
Exact match
Tasks
Multilingual academic QA
Authors
OpenAI (2026)
Source
MMMLU ↗

Scores are as reported by model developers or evaluators and may use different settings (for example reasoning effort or tools). Compare with care.