Compare AI models

Specifications
Parameters
8B16B
Context window
1M8k
Benchmarks
MATH
35.9%45.3%
MMLU
68.1%70%
Overview
CompanyGoogleMoonshot AI
Release dateOct 3 2024Feb 22 2025
AccessClosedOpen Weight
Model detailsView modelView model

Frequently asked questions

Moonlight leads Gemini 1.5 Flash-8B on 2 of the 2 benchmarks they both report (MATH, MMLU). Gemini 1.5 Flash-8B shipped 142 days before Moonlight, so benchmark comparisons should account for the intervening progress.

Gemini 1.5 Flash-8B has 8B parameters, while Moonlight has 16B. Context windows are 1M (Gemini 1.5 Flash-8B) vs 8k (Moonlight). Gemini 1.5 Flash-8B is closed, while Moonlight is open weight.

On MATH, Moonlight leads at 45.3% vs Gemini 1.5 Flash-8B at 35.9%. On MMLU, Moonlight leads at 70% vs Gemini 1.5 Flash-8B at 68.1%.