nvidia / text-generation updated 3 months ago

NVIDIA-Nemotron-3-Super-120B-A12B-FP8

Name: NVIDIA-Nemotron-3-Super-120B-A12B-FP8
Author: nvidia

:---:--- Total Parameters 120B (12B active) Architecture LatentMoE - Mamba-2 + MoE + Attention hybrid with Multi-Token Prediction (MTP) Context Length Up to 1M tokens Minimum GPU Requirement 2× H100-80GB Supported Languages English, French, German, Italian, Japanese, Spanish, Chinese Best For Agentic workflows, long-context reasoning, high-volume workloads (e.g. IT ticket automation), tool use, RAG Reasoning Mode Con...

Params

123.6 B

Context

262,144

Downloads 30d

208 K

Likes

271

Licence: other Not gated SAFETENSORS 7 languages View on Hugging Face ↗

Download history

daily snapshots · 10 days

225 K208 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	f8_e4m3	128.4 GB	160.2 GB	❌ Won’t fit
model.safetensors (bf16, full)	bf16 + 256K ctx	128.4 GB	735.0 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-super-120b-a12b-fp8
{
  "hf_id": "nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8",
  "params_b": 123.61,
  "context_length": 262144,
  "license": { "id": "other", "commercial": "unknown" },
  "downloads_30d": 208471,
  "vram_estimates": [
    { "quant": "f8_e4m3", "gb": 160.2 }
  ],
  "updated_at": "2026-07-28T18:07:39Z"
}

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

Specifications

Architecture: NemotronHForCausalLM
Parameters: 123.6 B
Tensor type: F8_E4M3
Context length: 262,144
Vocabulary: 131,072
Layers / heads: 88 / 32
Licence: other
First seen on the Hub: 2026-03-10
Training datasets: nvidia/nemotron-post-training-v3, nvidia/nemotron-pre-training-datasets
Added to our catalog: 2026-07-28

Compare with

Qwen3-32B 8.3 M dolphin-2.9.1-yi-1.5-34b 4.6 M gpt-oss-120b 4.2 M NVIDIA-Nemotron-3-Super-120B-A12B-NVFP4 2.8 M DeepSeek-V4-Flash 2.7 M Qwen3-Coder-Next-FP8 2.2 M

Appears in

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