Compare AI models

Specifications
Parameters
—16B
Context window
32k8k
Benchmarks
HumanEval
45.1%48.1%
WinoGrande
86.7%74.6%
MMLU
81.2%70%
GSM8K
81%77.4%
Overview
CompanyMistralMoonshot AI
Release dateFeb 26 2024Feb 22 2025
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Mistral Large leads Moonlight on 3 of the 4 benchmarks they both report (HumanEval, WinoGrande, MMLU, GSM8K). Mistral Large shipped 362 days before Moonlight, so benchmark comparisons should account for the intervening progress.

Context windows are 32k (Mistral Large) vs 8k (Moonlight). Mistral Large is closed, while Moonlight is open weight.

On HumanEval, Moonlight leads at 48.1% vs Mistral Large at 45.1%. On WinoGrande, Mistral Large leads at 86.7% vs Moonlight at 74.6%. On MMLU, Mistral Large leads at 81.2% vs Moonlight at 70%. On GSM8K, Mistral Large leads at 81% vs Moonlight at 77.4%.