Compare AI models

Specifications
Parameters
70B16B
Context window
16k8k
Benchmarks
HumanEval
53%48.1%
Overview
CompanyMetaMoonshot AI
Release dateJan 29 2024Feb 22 2025
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

Code Llama 70B leads Moonlight on 1 of the 1 benchmark they both report (HumanEval). Code Llama 70B shipped 390 days before Moonlight, so benchmark comparisons should account for the intervening progress.

Code Llama 70B has 70B parameters, while Moonlight has 16B. Context windows are 16k (Code Llama 70B) vs 8k (Moonlight).

On HumanEval, Code Llama 70B leads at 53% vs Moonlight at 48.1%.