Mistral-v0.3-6B: benchmark results

Provider: Mistral. Access: Open.

Unified ELO 1370 ± 20, rank #2351 of 2928 rated models, from 12 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open LLM Leaderboard - MuSR39.08Score35.5
Open LLM Leaderboard v1 - TruthfulQA MC245.64MC2 (%) (0-shot)31.6
Open LLM Leaderboard v1 - MMLU51.83Accuracy (%) (5-shot)31.1
Open LLM Leaderboard v1 - WinoGrande72.77Accuracy (%) (5-shot)30.1
Open LLM Leaderboard v1 - GSM8K8.34Accuracy (%) (5-shot)28.4
Open LLM Leaderboard - MMLU-Pro21.43Score22.5
Open LLM Leaderboard v1 - HellaSwag71.65Normalized accuracy (%) (10-shot)21.9
Open LLM Leaderboard - BBH37.74Score21.7
Open LLM Leaderboard v1 - ARC Challenge45.14Normalized accuracy (%) (25-shot)20.7
Open LLM Leaderboard - GPQA26.51Score19.3
Open LLM Leaderboard - IFEval24.54Score19.2
Open LLM Leaderboard - MATH Level 51.36Score9.9

Interactive version: theaggregate.ai/model?slug=mistral-v0-3-6b · How It Works · Data refreshed daily, snapshot 2026-09-23.