Mistral Small 4: benchmark results

Mistral AI's Small 4, a 119B sparse MoE (6.5B active) model (March 2026). Provider: Mistral. Released 2026-03-16. Access: Open.

Unified ELO 1551 ± 1, rank #428 of 1392 rated models, from 100 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
SpeechMap Compliance90.9% Requests Completed92.2
StereoTales117Emissions (self-reported)90.9
Guesswork 2026-070.93MAE (z-score units)75
UGI - Writing40.33Writing Score73.8
UGI - Natural Intelligence27.77NatInt Score68.7
BLXBench75.2Score (self-reported)66.7
Wolfram LLM Benchmarking Project43.1Correct Functionality (%)52.6
Conceptual Reasoning Index - Decision Theory (DTBench)51.55Chance-Corrected Score (0-100)51.3
SEA-HELM (Tamil)55.09Mean Score51.2
SEA-HELM (Filipino)56.18Mean Score48.8
AraGen33.853C3H Score (%)47.8
Epoch AI - Dtbench70.93Score46.9

Interactive version: theaggregate.ai/model?slug=mistral-small-4 · How It Works · Data refreshed daily, snapshot 2026-09-05.