Compare AI models

Specifications
Parameters
671B16B
Context window
128k8k
Benchmarks
LiveCodeBench
65.9%14.3%
GPQA Diamond
71.5%35.2%
SimpleQA
30.1%5.6%
MMLU-Pro
84%42.4%
MMLU
90.8%70%
Overview
CompanyDeepSeekMoonshot AI
Release dateJan 20 2025Feb 22 2025
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek R1 leads Moonlight on 5 of the 5 benchmarks they both report. DeepSeek R1 shipped 33 days before Moonlight, so benchmark comparisons should account for the intervening progress.

DeepSeek R1 has 671B parameters, while Moonlight has 16B. Context windows are 128k (DeepSeek R1) vs 8k (Moonlight).

On LiveCodeBench, DeepSeek R1 leads at 65.9% vs Moonlight at 14.3%. On GPQA Diamond, DeepSeek R1 leads at 71.5% vs Moonlight at 35.2%. On SimpleQA, DeepSeek R1 leads at 30.1% vs Moonlight at 5.6%. On MMLU-Pro, DeepSeek R1 leads at 84% vs Moonlight at 42.4%. On MMLU, DeepSeek R1 leads at 90.8% vs Moonlight at 70%.