Compare AI models

Specifications
Parameters
671B1B/4B/12B/27B
Context window
128k—
API pricing
Cheapest input
—$0.04DeepInfra
Cheapest output
—$0.08DeepInfra
—$15.00DeepInfra · bf16
Benchmarks
LiveCodeBench
65.9%29.7%
GPQA Diamond
71.5%42.4%
SimpleQA
30.1%10%
MMLU-Pro
84%67.5%
MMLU
90.8%76.9%
Overview
CompanyDeepSeekGoogle
Release dateJan 20 2025Mar 12 2025
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek R1 leads Gemma 3 on 5 of the 5 benchmarks they both report. DeepSeek R1 shipped 51 days before Gemma 3, so benchmark comparisons should account for the intervening progress.

DeepSeek R1 has 671B parameters, while Gemma 3 has 1B/4B/12B/27B.

On LiveCodeBench, DeepSeek R1 leads at 65.9% vs Gemma 3 at 29.7%. On GPQA Diamond, DeepSeek R1 leads at 71.5% vs Gemma 3 at 42.4%. On SimpleQA, DeepSeek R1 leads at 30.1% vs Gemma 3 at 10%. On MMLU-Pro, DeepSeek R1 leads at 84% vs Gemma 3 at 67.5%. On MMLU, DeepSeek R1 leads at 90.8% vs Gemma 3 at 76.9%.