What are the hardware requirements for DeepSeek V3.1 Int4 Mixed AutoRound?

DeepSeek V3.1 Int4 Mixed AutoRound requires approximately 225 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 V3.1 Int4 Mixed AutoRound and how large is it?

DeepSeek V3.1 Int4 Mixed AutoRound is developed by Intel. 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 V3.1 Int4 Mixed AutoRound?

DeepSeek V3.1 Int4 Mixed AutoRound is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

DeepSeek V3.1 Int4 Mixed AutoRound by Intel — VRAM 225GB, 160K context

Name: DeepSeek V3.1 Int4 Mixed AutoRound
Author: Intel

DeepSeek V3.1 Int4 Mixed AutoRound is an open-source language model by Intel. Features: LLM, VRAM: 225GB, Context: 160K, LLM Explorer Score: 0.19.

Arxiv:2309.05516 4-bit Auto-round Base model:deepseek-ai/deepsee... Base model:quantized:deepseek-... Conversational Custom code Deepseek v3 Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Intel/DeepSeek-V3.1-int4-mixed-AutoRound

DeepSeek V3.1 Int4 Mixed AutoRound Parameters and Internals

LLM Name	DeepSeek V3.1 Int4 Mixed AutoRound
Repository 🤗	https://huggingface.co/Intel/DeepSeek-V3.1-int4-mixed-AutoRound
Base Model(s)	deepseek-ai/DeepSeek-V3.1 deepseek-ai/DeepSeek-V3.1
Required VRAM	225 GB
Updated	2026-04-24
Maintainer	Intel
Model Type	deepseek_v3
Model Files	5.0 GB: 1-of-72 5.0 GB: 2-of-72 5.0 GB: 3-of-72 5.0 GB: 4-of-72 5.0 GB: 5-of-72 5.0 GB: 6-of-72 5.0 GB: 7-of-72 5.0 GB: 8-of-72 5.0 GB: 9-of-72 5.0 GB: 10-of-72 5.0 GB: 11-of-72 5.0 GB: 12-of-72 5.0 GB: 13-of-72 5.0 GB: 14-of-72 5.0 GB: 15-of-72 5.0 GB: 16-of-72 5.0 GB: 17-of-72 5.0 GB: 18-of-72 5.0 GB: 19-of-72 5.0 GB: 20-of-72 5.0 GB: 21-of-72 5.0 GB: 22-of-72 5.0 GB: 23-of-72 5.0 GB: 24-of-72 5.0 GB: 25-of-72 5.0 GB: 26-of-72 5.0 GB: 27-of-72 5.0 GB: 28-of-72 5.0 GB: 29-of-72 5.0 GB: 30-of-72 5.0 GB: 31-of-72 5.0 GB: 32-of-72 5.0 GB: 33-of-72 5.0 GB: 34-of-72 5.0 GB: 35-of-72 5.0 GB: 36-of-72 5.0 GB: 37-of-72 5.0 GB: 38-of-72 5.0 GB: 39-of-72 5.0 GB: 40-of-72 5.0 GB: 41-of-72 5.0 GB: 42-of-72 5.0 GB: 43-of-72 5.0 GB: 44-of-72 5.0 GB: 45-of-72
Model Architecture	DeepseekV3ForCausalLM
Context Length	163840
Model Max Length	163840
Transformers Version	4.56.0.dev0
Vocabulary Size	129280
Torch Data Type	bfloat16

Rank the DeepSeek V3.1 Int4 Mixed AutoRound 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 V3.1 Int4 Mixed AutoRound

Best Alternatives	Context / RAM	Downloads	Likes
Kimi K2 Thinking NVFP4	256K / 210 GB	15569	29
Kimi K2 Thinking BF16	256K / 220 GB	63	6
Kimi K2 Instruct 0905	256K / 667.4 GB	17	7
DeepSeek R1 0528	160K / 176.1 GB	48	15
DeepSeek R1 0528 FP4	160K / 107 GB	72991	36
DeepSeek V3.0324 FP4	160K / 242.8 GB	9984	12
DeepSeek R1 FP4	160K / 232 GB	6856	264
DeepSeek V3.1 MLX 5.5bit	160K / GB	400	5
...V3.1 Base Int4 Mixed AutoRound	160K / 220 GB	103	2
R1 1776	160K / 243.4 GB	22	9

Note: green Score (e.g. "73.2") means that the model is better than Intel/DeepSeek-V3.1-int4-mixed-AutoRound.