Compare AI models

Specifications
Parameters
671B8B/70B
Context window
128k8k
Benchmarks
MMLU
90.8%82%
Overview
CompanyDeepSeekMeta
Release dateJan 20 2025Apr 18 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek R1 leads LLaMA 3 (8B/70B) on 1 of the 1 benchmark they both report (MMLU). LLaMA 3 (8B/70B) shipped 277 days before DeepSeek R1, so benchmark comparisons should account for the intervening progress.

DeepSeek R1 has 671B parameters, while LLaMA 3 (8B/70B) has 8B/70B. Context windows are 128k (DeepSeek R1) vs 8k (LLaMA 3 (8B/70B)).

On MMLU, DeepSeek R1 leads at 90.8% vs LLaMA 3 (8B/70B) at 82%.