MMLU

Every result

• GPT-4 pretrained base (paper model)86.4%Reported by the model's maker (OpenAI), Mar 15, 2023

  OpenAI reported an 86.4% MMLU score for pretrained-base GPT-4 using five-shot prompting.

  Tasks: English-language, 57 subjects; exact split unspecified

  Limits: Provider-reported result, not independently reproduced or a served API snapshot. Exact dataset revision and score aggregation are unspecified here. Unknown run settings remain null; no unqualified cross-model rank is established. The report estimated limited training-data overlap using a separate contamination check; that check is not the score-run sample size.

  OpenAI; hosted by arXiv

• Mistral-7B-v0.160.1%Reported by the model's maker (Mistral AI / Jiang et al.), Oct 10, 2023

  Mistral AI reported a 60.1% MMLU score for Mistral 7B v0.1 base with five-shot evaluation.

  Tasks: Exact dataset split/revision unspecified in the inspected paper

  Limits: Provider-reported result from its own evaluation pipeline; not independently reproduced. This is the pretrained base, not Mistral 7B Instruct. Missing run metadata remain unknown; no cross-paper or universal model rank is established.

  Jiang et al. / Mistral AI; hosted by arXiv