nvidia / text-generation updated 5 months ago

NVIDIA-Nemotron-Nano-9B-v2

Name: NVIDIA-Nemotron-Nano-9B-v2
Author: nvidia

NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks. It responds to user queries and tasks by first generating a reasoning trace and then concluding with a final response. The model's reasoning capabilities can be controlled via a system prompt. If the user prefers the model to provide its final answer wi...

Params

8.9 B

Context

131,072

Downloads 30d

313 K

Likes

506

Licence: other Not gated SAFETENSORS 6 languages View on Hugging Face ↗

Download history

daily snapshots · 10 days

313 K299 K

Jul 28Jul 31Aug 3Aug 6

Can you run it?

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
model.safetensors	bf16	17.8 GB	21.4 GB	⚠️ Tight — reduce context
model.safetensors (bf16, full)	bf16 + 128K ctx	17.8 GB	41.4 GB	❌ Won’t fit

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Run it

copy-paste, exact tags checked against the Hub

~ · curl · api/v1

$ curl -s https://aimodelscomparison.com/api/v1/models/nvidia-nemotron-nano-9b-v2
{
  "hf_id": "nvidia/NVIDIA-Nemotron-Nano-9B-v2",
  "params_b": 8.89,
  "context_length": 131072,
  "license": { "id": "other", "commercial": "unknown" },
  "downloads_30d": 312725,
  "vram_estimates": [
    { "quant": "bf16", "gb": 21.4 }
  ],
  "updated_at": "2026-07-28T18:06:40Z"
}

est. VRAM —on RTX 4090 · 24 GBJSON API →

Specifications

Architecture: NemotronHForCausalLM
Parameters: 8.9 B
Tensor type: BF16
Context length: 131,072
Vocabulary: 131,072
Layers / heads: 56 / 40
Licence: other
First seen on the Hub: 2025-08-12
Training datasets: nvidia/Nemotron-Post-Training-Dataset-v1, nvidia/Nemotron-Post-Training-Dataset-v2, nvidia/Nemotron-Pretraining-Dataset-sample, nvidia/Nemotron-CC-v2, nvidia/Nemotron-CC-Math-v1, nvidia/Nemotron-Pretraining-SFT-v1
Added to our catalog: 2026-07-28

Family

Base model and the most-downloaded derivatives in the catalog.

Derivative NVIDIA-Nemotron-3-Nano-4B-BF16 nvidia 739 K

Compare with

Qwen3-8B 16.3 M Qwen2.5-7B-Instruct 12.1 M Qwen3.6-35B-A3B-NVFP4 11.3 M Llama-3.1-8B-Instruct 8.0 M Qwen2.5-7B-Instruct-AWQ 4.6 M granite-4.1-8b 4.2 M

Appears in

Best open coding models Commercial-use text models Models that fit in 12 GB