Compare AI models

Specifications
Parameters
671B9B/27B
Context window
128k8k
API pricing
Cheapest input
—$0.65NextBit
Cheapest output
—$0.65NextBit
—$163NextBit · int4
Benchmarks
MMLU
90.8%75.2%
Overview
CompanyDeepSeekGoogle
Release dateJan 20 2025Jun 27 2024
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek R1 leads Gemma 2 on 1 of the 1 benchmark they both report (MMLU). Gemma 2 shipped 207 days before DeepSeek R1, so benchmark comparisons should account for the intervening progress.

DeepSeek R1 has 671B parameters, while Gemma 2 has 9B/27B. Context windows are 128k (DeepSeek R1) vs 8k (Gemma 2).

On MMLU, DeepSeek R1 leads at 90.8% vs Gemma 2 at 75.2%.