Compare AI models

Specifications
Parameters
671B2B/7B
Context window
128k8k
Benchmarks
MMLU
90.8%64.3%
Overview
CompanyDeepSeekGoogle
Release dateJan 20 2025Feb 21 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek R1 leads Gemma 1 on 1 of the 1 benchmark they both report (MMLU). Gemma 1 shipped 334 days before DeepSeek R1, so benchmark comparisons should account for the intervening progress.

DeepSeek R1 has 671B parameters, while Gemma 1 has 2B/7B. Context windows are 128k (DeepSeek R1) vs 8k (Gemma 1).

On MMLU, DeepSeek R1 leads at 90.8% vs Gemma 1 at 64.3%.