Compare AI models

Specifications
Parameters
671B16B
Context window
128k8k
API pricing
Cheapest input
$0.256DeepInfra—
Cheapest output
$0.712DeepInfra—
$103StreamLake—
Benchmarks
LiveCodeBench
40.5%14.3%
GPQA Diamond
59.1%35.2%
MMLU
88.5%70%
Overview
CompanyDeepSeekMoonshot AI
Release dateDec 26 2024Feb 22 2025
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V3 leads Moonlight on 3 of the 3 benchmarks they both report (LiveCodeBench, GPQA Diamond, MMLU). DeepSeek V3 shipped 58 days before Moonlight, so benchmark comparisons should account for the intervening progress.

DeepSeek V3 has 671B parameters, while Moonlight has 16B. Context windows are 128k (DeepSeek V3) vs 8k (Moonlight).

On LiveCodeBench, DeepSeek V3 leads at 40.5% vs Moonlight at 14.3%. On GPQA Diamond, DeepSeek V3 leads at 59.1% vs Moonlight at 35.2%. On MMLU, DeepSeek V3 leads at 88.5% vs Moonlight at 70%.