What are the hardware requirements for DeepSeek R1 NVFP4?

DeepSeek R1 NVFP4 requires approximately 232 GB of VRAM and supports a context window of 160K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed DeepSeek R1 NVFP4 and how large is it?

DeepSeek R1 NVFP4 is developed by nvidia, a model with 396.8b 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 DeepSeek R1 NVFP4?

DeepSeek R1 NVFP4 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

DeepSeek R1 NVFP4 by nvidia — VRAM 232GB, 160K context

Name: DeepSeek R1 NVFP4
Author: nvidia

DeepSeek R1 NVFP4 is an open-source language model by nvidia. Features: 396.8b LLM, VRAM: 232GB, Context: 160K, License: mit, LLM Explorer Score: 0.21.

8-bit Base model:deepseek-ai/deepsee... Base model:finetune:deepseek-a... Conversational Custom code Deepseek v3 Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/nvidia/DeepSeek-R1-NVFP4

DeepSeek R1 NVFP4 Parameters and Internals

LLM Name	DeepSeek R1 NVFP4
Repository 🤗	https://huggingface.co/nvidia/DeepSeek-R1-NVFP4
Base Model(s)	DeepSeek R1 deepseek-ai/DeepSeek-R1
Model Size	396.8b
Required VRAM	232 GB
Updated	2026-08-03
Maintainer	nvidia
Model Type	deepseek_v3
Model Files	5.4 GB: 1-of-80 5.4 GB: 2-of-80 5.4 GB: 3-of-80 5.4 GB: 4-of-80 5.4 GB: 5-of-80 5.4 GB: 6-of-80 5.4 GB: 7-of-80 5.4 GB: 8-of-80 5.4 GB: 9-of-80 5.4 GB: 10-of-80 5.4 GB: 11-of-80 5.4 GB: 12-of-80 5.4 GB: 13-of-80 5.4 GB: 14-of-80 5.4 GB: 15-of-80 5.4 GB: 16-of-80 5.4 GB: 17-of-80 5.4 GB: 18-of-80 5.4 GB: 19-of-80 5.2 GB: 20-of-80 5.4 GB: 21-of-80 5.4 GB: 22-of-80 5.4 GB: 23-of-80 5.4 GB: 24-of-80 5.4 GB: 25-of-80 5.4 GB: 26-of-80 5.4 GB: 27-of-80 5.4 GB: 28-of-80 5.4 GB: 29-of-80 5.4 GB: 30-of-80 5.4 GB: 31-of-80 5.4 GB: 32-of-80 5.4 GB: 33-of-80 5.4 GB: 34-of-80 5.4 GB: 35-of-80 5.4 GB: 36-of-80 5.4 GB: 37-of-80 5.4 GB: 38-of-80 5.4 GB: 39-of-80 5.4 GB: 40-of-80 5.4 GB: 41-of-80 5.4 GB: 42-of-80 5.4 GB: 43-of-80
Model Architecture	DeepseekV3ForCausalLM
License	mit
Context Length	163840
Model Max Length	163840
Transformers Version	4.46.3
Vocabulary Size	129280
Torch Data Type	bfloat16

Rank the DeepSeek R1 NVFP4 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 DeepSeek R1 NVFP4

Best Alternatives	Context / RAM	Downloads	Likes
DeepSeek V3.0324 NVFP4	160K / 242.8 GB	34174	17
DeepSeek R1 0528 NVFP4	160K / 107 GB	3174	44

Note: green Score (e.g. "73.2") means that the model is better than nvidia/DeepSeek-R1-NVFP4.