Nemotron-Nano-3-30B-A3B is a large language model trained by NVIDIA, designed as a unified model for both reasoning and non-reasoning tasks. It employs a hybrid Mamba-2 + MoE architecture with 30B total parameters and 3.5B active parameters. Supports English, German, Spanish, French, Italian, and Japanese.
On-demand DeploymentDocs | On-demand deployments allow you to use NVIDIA Nemotron Nano 3 30B A3B on dedicated GPUs with Fireworks' high-performance serving stack with high reliability and no rate limits. |
