Skip to content

Nemotron 3

by NVIDIA · Website

NVIDIA's Nemotron 3 family features novel hybrid architectures combining Mamba and Transformer blocks. Optimized for inference efficiency with strong performance on coding, math, and reasoning tasks.

Variants (1)

Nemotron 3 Nano 8B

8B

NVIDIA

Min 7.5 GB
text-generation code-generation reasoning