mpt-30B — benchmark results

Provider: Databricks. Released 2023-06-22. Access: Open.

Unified ELO 1272 ± 8, rank #1585 of 1776 rated models, from 98 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
HELM Classic - Entity Data Imputation85.52Exact Match (%)100
OpenEval - IMDb95.74Exact Match (%)100
HELM Classic - LSAT24.78Exact Match (%)97.1
HELM Classic - WikiFact36.93Exact Match (%)92.4
HELM Classic - Entity Matching89.74Exact Match (%)89.4
HELM Classic - RAFT72.27Exact Match (%)89.4
MMLU-by-task - Machine Learning45.54Accuracy (%)89.4
HELM Classic - IMDB95.9Exact Match (%)87.9
HELM Classic - NarrativeQA73.15F1 (%)86.2
HELM Classic - bAbI54.97Exact Match (%)81.9
HELM Classic - NaturalQuestions Open Book67.29F1 (%)81.5
MMLU-by-task - College Mathematics36Accuracy (%)80.3

Interactive version: theaggregate.ai/model?slug=mpt-30b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.