Mistral-NeMo-Minitron-8B-Instruct — benchmark results

NVIDIA's instruct-aligned version of Mistral-NeMo-Minitron-8B-Base, tuned with NeMo Aligner for chat, RAG, and function calling (8K context). Provider: Mistral. Released 2024-10-02. Access: Open.

Unified ELO 1453 ± 14, rank #982 of 1776 rated models, from 62 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
LatamBoard - Spanish PAWS63Score (%)86.4
EuroEval Spanish NLU - Sentiment Headlines ES48.9Sentiment classification Score (%)85
LatamBoard - Spanish OpenBookQA38Score (%)84.8
LatamBoard - BLUEX61.34Score (%)81.8
LatamBoard - ENEM Challenge73.41Score (%)81.8
LatamBoard - Spanish COPA83.6Score (%)81.8
LatamBoard - Spanish XNLI47.83Score (%)78.8
LatamBoard - Spanish MGSM12.4Score (%)77.3
LatamBoard - Spanish Score62.02Score (%)75.8
Open LLM Leaderboard - MMLU-Pro33.23Score73.5
LatamBoard - OAB Exams51.3Score (%)72.7
EuroEval Spanish NLU - ScaLA ES28.35Linguistic acceptability Score (%)71.4

Interactive version: theaggregate.ai/model?slug=mistral-nemo-minitron-8b-instruct · How the rankings work · Data refreshed daily, snapshot 2026-07-22.