Model Detail

N-ATLaS is an AI model with 4.0B parameters released by NCAIR1. It has accumulated 6K downloads on Hugging Face since publication.

N-ATLaS ships with 4.0B parameters. Total weight footprint is approximately 8.0 GB, which is the relevant figure when planning local-inference VRAM. Access is gated on Hugging Face — a manual approval step is required before weights can be downloaded.

N-ATLaS is best fit for general-purpose AI workloads. Treat this as a starting matrix rather than a benchmark verdict — the right deployment usually depends on the specific evaluation suite that mirrors your workload.