Compare AI models

Specifications
Parameters
141B16B
Context window
64k8k
API pricing
Cheapest input
$2.00Mistral—
Cheapest output
$6.00Mistral—
$700Mistral · zdr—
Benchmarks
HumanEval
45.1%48.1%
WinoGrande
84.7%74.6%
MMLU
77.75%70%
GSM8K
78.6%77.4%
Overview
CompanyMistralMoonshot AI
Release dateApr 10 2024Feb 22 2025
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Mixtral 8×22B leads Moonlight on 3 of the 4 benchmarks they both report (HumanEval, WinoGrande, MMLU, GSM8K). Mixtral 8×22B shipped 318 days before Moonlight, so benchmark comparisons should account for the intervening progress.

Mixtral 8×22B has 141B parameters, while Moonlight has 16B. Context windows are 64k (Mixtral 8×22B) vs 8k (Moonlight).

On HumanEval, Moonlight leads at 48.1% vs Mixtral 8×22B at 45.1%. On WinoGrande, Mixtral 8×22B leads at 84.7% vs Moonlight at 74.6%. On MMLU, Mixtral 8×22B leads at 77.75% vs Moonlight at 70%. On GSM8K, Mixtral 8×22B leads at 78.6% vs Moonlight at 77.4%.