Ministral 3 14B — benchmark results
Mistral's Apache-2.0 14B dense model from the Mistral 3 family, with image understanding and single-GPU focus (December 2025). Provider: Mistral. Released 2025-12-02. Access: Open.
Unified ELO 1545 ± 7, rank #638 of 1776 rated models, from 595 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval English Summarization - CNN Dailymail | 46.05 | Score (%) | 100 |
| EuroEval Finnish NLU - Tydiqa FI | 78.61 | Reading comprehension Score (%) | 100 |
| Konkur 1404 - Foreign Language | 44.23 | Accuracy (%, text-only) | 100 |
| EuroEval Italian NLU - SQuAD IT | 78.66 | Reading comprehension Score (%) | 99.8 |
| EuroEval Dutch NLU - SQuAD NL | 86.09 | Reading comprehension Score (%) | 99.7 |
| EuroEval French NLU - Allocine | 97.8 | Sentiment classification Score (%) | 99.7 |
| EuroEval Icelandic NLU - NQII | 65.72 | Reading comprehension Score (%) | 99.7 |
| EuroEval Danish NLU - Multi Wiki QA DA | 82.94 | Reading comprehension Score (%) | 99.6 |
| EuroEval Estonian NLU - MultiWikiQA ET | 77.94 | Reading comprehension Score (%) | 99.6 |
| EuroEval Lithuanian NLU - MultiWikiQA LT | 76.12 | Reading comprehension Score (%) | 99.6 |
| EuroEval Polish NLU - Polemo2 | 98.2 | Sentiment classification Score (%) | 99.6 |
| EuroEval Faroese NLU - FoQA | 79.94 | Reading comprehension Score (%) | 99.5 |
Interactive version: theaggregate.ai/model?slug=ministral-3-14b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.