Compare AI models

Specifications
Parameters
12B16B
Context window
128k8k
Benchmarks
HumanEval
72%48.1%
MATH
48.1%45.3%
MMLU
69.2%70%
Overview
CompanyMistralMoonshot AI
Release dateSep 11 2024Feb 22 2025
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Pixtral 24.09 leads Moonlight on 2 of the 3 benchmarks they both report (HumanEval, MATH, MMLU). Pixtral 24.09 shipped 164 days before Moonlight, so benchmark comparisons should account for the intervening progress.

Pixtral 24.09 has 12B parameters, while Moonlight has 16B. Context windows are 128k (Pixtral 24.09) vs 8k (Moonlight).

On HumanEval, Pixtral 24.09 leads at 72% vs Moonlight at 48.1%. On MATH, Pixtral 24.09 leads at 48.1% vs Moonlight at 45.3%. On MMLU, Moonlight leads at 70% vs Pixtral 24.09 at 69.2%.