What are the hardware requirements for NVIDIA Nemotron 3 Super 120B A12B FP8?

NVIDIA Nemotron 3 Super 120B A12B FP8 requires approximately 128.4 GB of VRAM and supports a context window of 256K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed NVIDIA Nemotron 3 Super 120B A12B FP8 and how large is it?

NVIDIA Nemotron 3 Super 120B A12B FP8 is developed by nvidia, a model with 120b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate NVIDIA Nemotron 3 Super 120B A12B FP8?

NVIDIA Nemotron 3 Super 120B A12B FP8 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

NVIDIA Nemotron 3 Super 120B A12B FP8 by nvidia — VRAM 128.4GB, 256K context

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

NVIDIA Nemotron 3 Super 120B A12B FP8 is an open-source language model by nvidia. Features: 120b LLM, VRAM: 128.4GB, Context: 256K, License: other, LLM Explorer Score: 0.36.

Arxiv:2512.20848 Arxiv:2512.20856 Conversational Custom code Dataset:nvidia/nemotron-post-t... Dataset:nvidia/nemotron-pre-tr... De Deploy:azure Deploy:sagemaker En Endpoints compatible Es Eval-results Fr It Ja Latent-moe Modelopt Mtp Nemotron-3 Nemotron h Nvidia Pytorch Region:us Safetensors Sharded Tensorflow Zh

Model Card on HF 🤗: https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8

NVIDIA Nemotron 3 Super 120B A12B FP8 Parameters and Internals

LLM Name	NVIDIA Nemotron 3 Super 120B A12B FP8
Repository 🤗	https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8
Model Size	120b
Required VRAM	128.4 GB
Updated	2026-07-31
Maintainer	nvidia
Model Type	nemotron_h
Model Files	5.0 GB: 1-of-26 5.0 GB: 2-of-26 5.0 GB: 3-of-26 5.0 GB: 4-of-26 5.0 GB: 5-of-26 5.0 GB: 6-of-26 5.0 GB: 7-of-26 5.0 GB: 8-of-26 5.0 GB: 9-of-26 5.0 GB: 10-of-26 5.0 GB: 11-of-26 5.0 GB: 12-of-26 5.0 GB: 13-of-26 5.0 GB: 14-of-26 5.0 GB: 15-of-26 5.0 GB: 16-of-26 5.0 GB: 17-of-26 5.0 GB: 18-of-26 5.0 GB: 19-of-26 5.0 GB: 20-of-26 5.0 GB: 21-of-26 5.0 GB: 22-of-26 5.0 GB: 23-of-26 5.0 GB: 24-of-26 5.0 GB: 25-of-26 3.4 GB: 26-of-26
Supported Languages	en fr es it de ja zh
Model Architecture	NemotronHForCausalLM
License	other
Context Length	262144
Model Max Length	262144
Transformers Version	4.57.6
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|im_end\|>
Vocabulary Size	131072

Rank the NVIDIA Nemotron 3 Super 120B A12B FP8 capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to NVIDIA Nemotron 3 Super 120B A12B FP8

Best Alternatives	Context / RAM	Downloads	Likes
...on 3 Super 120B A12B Base BF16	1024K / 209.5 GB	13053	32
...motron 3 Super 120B A12B NVFP4	256K / 80.3 GB	3041935	408
...motron 3 Super 120B A12B NVFP4	256K / 80.3 GB	1132	27
...emotron 3 Super 120B A12B BF16	256K / 194.6 GB	864394	414
...DIA Nemotron 3 Super 120B A12B	256K / 214.5 GB	294	3
...Nemotron 3 Super 120B A12B FP8	256K / 128.4 GB	81	10
... Super 64B A12B Math REAP BF16	256K / 128.6 GB	1814	1
Kenosistron Bf16	256K / 199.6 GB	29	0
...20B A12B BF16 REAP 50pct Draft	256K / 128.5 GB	318	6
...n 3 Super 120B A12B BF16 MTPv2	256K / 5.9 GB	5	6

Note: green Score (e.g. "73.2") means that the model is better than nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8.