Compare AI models

Specifications
Parameters
671B16B
Context window
128k8k
API pricing
Cheapest input
$0.27SiliconFlow—
Cheapest output
$0.41SiliconFlow—
$74.50SiliconFlow · fp8—
Benchmarks
LiveCodeBench
74.1%14.3%
GPQA Diamond
79.9%35.2%
MMLU-Pro
85%42.4%
Overview
CompanyDeepSeekMoonshot AI
Release dateSep 29 2025Feb 22 2025
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V3.2 Exp leads Moonlight on 3 of the 3 benchmarks they both report (LiveCodeBench, GPQA Diamond, MMLU-Pro). Moonlight shipped 219 days before DeepSeek V3.2 Exp, so benchmark comparisons should account for the intervening progress.

DeepSeek V3.2 Exp has 671B parameters, while Moonlight has 16B. Context windows are 128k (DeepSeek V3.2 Exp) vs 8k (Moonlight).

On LiveCodeBench, DeepSeek V3.2 Exp leads at 74.1% vs Moonlight at 14.3%. On GPQA Diamond, DeepSeek V3.2 Exp leads at 79.9% vs Moonlight at 35.2%. On MMLU-Pro, DeepSeek V3.2 Exp leads at 85% vs Moonlight at 42.4%.