AI Hub
All models
LLMOpen weights

Nemotron 3 Nano

NVIDIA

Updated May 21, 2026 · Benchmarks via Artificial Analysis, specs & pricing via OpenRouter · Methodology · Report an error

Specifications

Type
LLM
Access
Open weights
Released
December 15, 2025
License
NVIDIA Open Model License
Parameters
31.6B (3.6B active)
Context window
1,000,000 tokens
Input
Text
Output
Text

Our take

First of the Nemotron 3 family — a hybrid Mamba-2/Transformer MoE with a 1M context and large throughput gains over comparably sized models.

Links

Compare Nemotron 3 Nano with

See all Nemotron 3 Nano alternatives →