nvidia / text-generation updated 4 months ago

NVIDIA-Nemotron-3-Nano-4B-BF16

Name: NVIDIA-Nemotron-3-Nano-4B-BF16
Author: nvidia

NVIDIA-Nemotron-3-Nano-4B-BF16 is a small language model (SLM) 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 answe...

Params

4.0 B

Context

262,144

Downloads 30d

739 K

Likes

105

Licence: other Not gated SAFETENSORS 1 languages View on Hugging Face ↗

Download history

daily snapshots · 10 days

739 K710 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	7.9 GB	9.8 GB	✅ Runs comfortably
model.safetensors (bf16, full)	bf16 + 256K ctx	7.9 GB	28.3 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-3-nano-4b-bf16
{
  "hf_id": "nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16",
  "params_b": 3.97,
  "context_length": 262144,
  "license": { "id": "other", "commercial": "unknown" },
  "downloads_30d": 739025,
  "vram_estimates": [
    { "quant": "bf16", "gb": 9.8 }
  ],
  "updated_at": "2026-07-28T18:04:16Z"
}

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

Specifications

Architecture: NemotronHForCausalLM
Parameters: 4.0 B
Tensor type: BF16
Context length: 262,144
Vocabulary: 131,072
Layers / heads: 42 / 40
Licence: other
First seen on the Hub: 2026-03-07
Base model: NVIDIA-Nemotron-Nano-9B-v2
Training datasets: nvidia/Nemotron-CC-v2, nvidia/Nemotron-Post-Training-Dataset-v2, nvidia/Nemotron-Science-v1, nvidia/Nemotron-Instruction-Following-Chat-v1, nvidia/Nemotron-Agentic-v1, nvidia/Nemotron-Competitive-Programming-v1, nvidia/Nemotron-Math-Proofs-v1, nvidia/Nemotron-RL-Agentic-Conversational-Tool-Use-Pivot-v1, nvidia/Nemotron-RL-instruction_following, nvidia/Nemotron-RL-agent-calendar_scheduling, nvidia/Nemotron-RL-instruction_following-structured_outputs
Added to our catalog: 2026-07-28

Family

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

Base NVIDIA-Nemotron-Nano-9B-v2 nvidia 313 K

Compare with

Qwen3-8B 16.3 M Qwen2.5-1.5B-Instruct 14.0 M Qwen2.5-7B-Instruct 12.1 M Qwen3-1.7B 7.8 M Qwen2.5-3B-Instruct 6.0 M Qwen3-4B 4.6 M

Appears in

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