Compare AI models

Specifications
Parameters
671B8B/70B/405B
Context window
128k128k
API pricing
Cheapest input
$0.256DeepInfra$0.02DeepInfra
Cheapest output
$0.712DeepInfra$0.04DeepInfra
$103StreamLake$6.00DeepInfra · fp8
Benchmarks
GPQA Diamond
59.1%49%
MMLU
88.5%87.3%
Overview
CompanyDeepSeekMeta
Release dateDec 26 2024Jul 23 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek V3 leads LLaMA 3.1 on 2 of the 2 benchmarks they both report (GPQA Diamond, MMLU). LLaMA 3.1 shipped 156 days before DeepSeek V3, so benchmark comparisons should account for the intervening progress.

DeepSeek V3 has 671B parameters, while LLaMA 3.1 has 8B/70B/405B. Context windows are 128k (DeepSeek V3) vs 128k (LLaMA 3.1).

On GPQA Diamond, DeepSeek V3 leads at 59.1% vs LLaMA 3.1 at 49%. On MMLU, DeepSeek V3 leads at 88.5% vs LLaMA 3.1 at 87.3%.