What are the hardware requirements for Smaug Nemotron 3.5?

Smaug Nemotron 3.5 requires approximately 156 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Smaug Nemotron 3.5 and how large is it?

Smaug Nemotron 3.5 is developed by abacusai, 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 Smaug Nemotron 3.5?

Smaug Nemotron 3.5 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Smaug Nemotron 3.5 by abacusai — VRAM 156GB

Name: Smaug Nemotron 3.5
Author: abacusai

Smaug Nemotron 3.5 is an open-source language model by abacusai. Features: 120b LLM, VRAM: 156GB, License: other, LLM Explorer Score: 0.27.

Abacusai Agentic Base model:finetune:nvidia/nvi... Base model:nvidia/nvidia-nemot... Conversational Custom code En Endpoints compatible Image-text-to-text Nemotron-3.5 Nemotron h omni Region:us Safetensors Sharded Smaug Tensorflow

Model Card on HF 🤗: https://huggingface.co/abacusai/Smaug-Nemotron-3.5

Smaug Nemotron 3.5 Parameters and Internals

LLM Name	Smaug Nemotron 3.5
Repository 🤗	https://huggingface.co/abacusai/Smaug-Nemotron-3.5
Base Model(s)	nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16
Model Size	120b
Required VRAM	156 GB
Updated	2026-10-26
Maintainer	abacusai
Model Type	nemotron_h_omni
Model Files	4.0 GB: 1-of-63 4.0 GB: 2-of-63 4.0 GB: 3-of-63 4.0 GB: 4-of-63 4.0 GB: 5-of-63 4.0 GB: 6-of-63 4.0 GB: 7-of-63 4.0 GB: 8-of-63 4.0 GB: 9-of-63 4.0 GB: 10-of-63 4.0 GB: 11-of-63 4.0 GB: 12-of-63 4.0 GB: 13-of-63 4.0 GB: 14-of-63 4.0 GB: 15-of-63 4.0 GB: 16-of-63 4.0 GB: 17-of-63 4.0 GB: 18-of-63 4.0 GB: 19-of-63 4.0 GB: 20-of-63 4.0 GB: 21-of-63 4.0 GB: 22-of-63 4.0 GB: 23-of-63 4.0 GB: 24-of-63 4.0 GB: 25-of-63 4.0 GB: 26-of-63 4.0 GB: 27-of-63 4.0 GB: 28-of-63 4.0 GB: 29-of-63 4.0 GB: 30-of-63 4.0 GB: 31-of-63 4.0 GB: 32-of-63 4.0 GB: 33-of-63 4.0 GB: 34-of-63 4.0 GB: 35-of-63 4.0 GB: 36-of-63 4.0 GB: 37-of-63 4.0 GB: 38-of-63 4.0 GB: 39-of-63
Supported Languages	en
Model Architecture	NemotronH_Omni_Reasoning_V3
License	other
Transformers Version	5.12.1

Rank the Smaug Nemotron 3.5 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 Smaug Nemotron 3.5

Best Alternatives	Context / RAM	Downloads	Likes
...n 3.5 Super VL 120B A12B NVFP4	0K / 30 GB	9322	169
...on 3.5 Super VL 120B A12B BF16	0K / 148 GB	15095	78
...5 Super VL 120B A12B Base BF16	0K / 144 GB	1030	51
...Nemotron 3.5 Bolt 74B A9B BF16	0K / 82.8 GB	2193	40
...n 3.5 Super VL 120B A12B NVFP4	0K / 31.7 GB	1256	3
Smaug Nemotron 3.5 NVFP4	0K / 31.7 GB	81	5
...emotron 3.5 Super VL 120B A12B	0K / 160 GB	406	5
...n 3.5 Super VL 120B A12B NVFP4	0K / 78.6 GB	448	1
....5 Super VL 120B A12B EXL3 2.0	0K / 36.6 GB	217	1
...uper VL 120B A12B W4A16 Merlin	0K / 37.1 GB	48	2

Note: green Score (e.g. "73.2") means that the model is better than abacusai/Smaug-Nemotron-3.5.