Compare AI models

Specifications
Parameters
16B—
Context window
8k128k
Benchmarks
HumanEval
48.1%88.6%
GPQA Diamond
35.2%69.5%
SimpleQA
5.6%62.5%
MMLU
70%90.8%
Overview
CompanyMoonshot AIOpenAI
Release dateFeb 22 2025Feb 27 2025
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

GPT-4.5 leads Moonlight on 4 of the 4 benchmarks they both report (HumanEval, GPQA Diamond, SimpleQA, MMLU). Moonlight shipped 5 days before GPT-4.5, so benchmark comparisons should account for the intervening progress.

Context windows are 8k (Moonlight) vs 128k (GPT-4.5). Moonlight is open weight, while GPT-4.5 is closed.

On HumanEval, GPT-4.5 leads at 88.6% vs Moonlight at 48.1%. On GPQA Diamond, GPT-4.5 leads at 69.5% vs Moonlight at 35.2%. On SimpleQA, GPT-4.5 leads at 62.5% vs Moonlight at 5.6%. On MMLU, GPT-4.5 leads at 90.8% vs Moonlight at 70%.