Ministral-3-14B-Reasoning-2512 — benchmark results

Mistral Ministral 3 14B Reasoning 2512 snapshot. Provider: Mistral. Released 2025-12-01. Access: Open.

Unified ELO 1471 ± 21, rank #905 of 1776 rated models, from 294 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
EuroEval Bosnian NLU66NLU Average Score (%)100
EuroEval Croatian NLU - Multi Wiki QA HR73.9Reading comprehension Score (%)100
EuroEval Estonian NLU - MultiWikiQA ET77.98Reading comprehension Score (%)100
EuroEval Serbian NLU - MultiWikiQA SR68.07Reading comprehension Score (%)100
EuroEval Ukrainian NLU - MultiWikiQA UK71.57Reading comprehension Score (%)100
EuroEval Bulgarian NLU - MultiWikiQA BG80.66Reading comprehension Score (%)99.5
EuroEval French NLU - ScaLA FR73.6Linguistic acceptability Score (%)99.5
EuroEval Italian NLU - SQuAD IT78.42Reading comprehension Score (%)99.5
EuroEval Czech NLU - SQAD72.74Reading comprehension Score (%)99.1
EuroEval Portuguese NLU - MultiWikiQA PT79.73Reading comprehension Score (%)99.1
EuroEval Albanian NLU - MultiWikiQA SQ70.23Reading comprehension Score (%)99
EuroEval Hungarian NLU - MultiWikiQA HU71.83Reading comprehension Score (%)99

Interactive version: theaggregate.ai/model?slug=ministral-3-14b-reasoning-2512 · How the rankings work · Data refreshed daily, snapshot 2026-07-22.