Compare AI models

Specifications
Parameters
671B—
Context window
128k200k
Benchmarks
MATH-500
97.3%97.3%
DROP
92.279.2
GPQA Diamond
71.5%74.9%
SimpleQA
30.1%13.4%
MMLU
90.8%85.9%
Overview
CompanyDeepSeekOpenAI
Release dateJan 20 2025Jan 31 2025
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

DeepSeek R1 leads o3-mini on 3 of the 5 benchmarks they both report. DeepSeek R1 shipped 11 days before o3-mini, so benchmark comparisons should account for the intervening progress.

Context windows are 128k (DeepSeek R1) vs 200k (o3-mini). DeepSeek R1 is open weight, while o3-mini is closed.

On MATH-500, both models score 97.3%. On DROP, DeepSeek R1 leads at 92.2 vs o3-mini at 79.2. On GPQA Diamond, o3-mini leads at 74.9% vs DeepSeek R1 at 71.5%. On SimpleQA, DeepSeek R1 leads at 30.1% vs o3-mini at 13.4%. On MMLU, DeepSeek R1 leads at 90.8% vs o3-mini at 85.9%.