Compare AI models

Specifications
Parameters
1.5B/7B/8B/14B/32B/70B4B
Context window
—256k
Benchmarks
MATH-500
94.5%95.4%
GPQA Diamond
65.2%53.2%
Overview
CompanyDeepSeekNVIDIA
Release dateJan 20 2025Mar 16 2026
AccessOpen WeightOpen Weight
Model detailsView modelView model

Frequently asked questions

DeepSeek R1 Distill and Nemotron 3 Nano 4B are evenly matched across the 2 benchmarks they both report (MATH-500, GPQA Diamond). DeepSeek R1 Distill shipped 420 days before Nemotron 3 Nano 4B, so benchmark comparisons should account for the intervening progress.

DeepSeek R1 Distill has 1.5B/7B/8B/14B/32B/70B parameters, while Nemotron 3 Nano 4B has 4B.

On MATH-500, Nemotron 3 Nano 4B leads at 95.4% vs DeepSeek R1 Distill at 94.5%. On GPQA Diamond, DeepSeek R1 Distill leads at 65.2% vs Nemotron 3 Nano 4B at 53.2%.