Mistral 7B Instruct v0.3 — Benchmarks

Benchmark scores for Mistral 7B Instruct v0.3 aggregated from public leaderboards, with how it ranks among open models. See hardware requirements for what you need to run it.

Overall rank: #73 of 73 open modelscomposite 16.7/100 across 3 benchmarks in 3 categories · methodology

Knowledge

BenchmarkScoreOpen rankAll models
MMLU59.9%#44 / 76#91 / 136

Math

BenchmarkScoreOpen rankAll models
MATH Level 53.6%#31 / 32#107 / 108

Reasoning

BenchmarkScoreOpen rankAll models
GPQA Diamond15.2%#45 / 46#180 / 182

Scores aggregated from public benchmark sources (each linked from the benchmark pages). llmrun does not run these benchmarks.