What are the hardware requirements for DeepSeek R1 AWQ?

DeepSeek R1 AWQ 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 R1 AWQ and how large is it?

DeepSeek R1 AWQ is developed by cognitivecomputations. 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 AWQ?

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

DeepSeek R1 AWQ by cognitivecomputations — VRAM 225GB, 160K context

Name: DeepSeek R1 AWQ
Author: cognitivecomputations

DeepSeek R1 AWQ is an open-source language model by cognitivecomputations. Features: LLM, VRAM: 225GB, Context: 160K, License: mit, Quantized, LLM Explorer Score: 0.19.

4-bit Autotrain compatible Awq Base model:deepseek-ai/deepsee... Base model:quantized:deepseek-... Conversational Custom code Deepseek v3 En Endpoints compatible Quantized Region:us Safetensors Sharded Tensorflow Zh

Model Card on HF 🤗: https://huggingface.co/cognitivecomputations/DeepSeek-R1-AWQ

DeepSeek R1 AWQ Parameters and Internals

LLM Name	DeepSeek R1 AWQ
Repository 🤗	https://huggingface.co/cognitivecomputations/DeepSeek-R1-AWQ
Base Model(s)	DeepSeek R1 deepseek-ai/DeepSeek-R1
Required VRAM	225 GB
Updated	2025-07-13
Maintainer	cognitivecomputations
Model Type	deepseek_v3
Model Files	5.0 GB: 1-of-74 5.0 GB: 2-of-74 5.0 GB: 3-of-74 5.0 GB: 4-of-74 5.0 GB: 5-of-74 5.0 GB: 6-of-74 5.0 GB: 7-of-74 5.0 GB: 8-of-74 5.0 GB: 9-of-74 5.0 GB: 10-of-74 5.0 GB: 11-of-74 5.0 GB: 12-of-74 5.0 GB: 13-of-74 5.0 GB: 14-of-74 5.0 GB: 15-of-74 5.0 GB: 16-of-74 5.0 GB: 17-of-74 5.0 GB: 18-of-74 5.0 GB: 19-of-74 5.0 GB: 20-of-74 5.0 GB: 21-of-74 5.0 GB: 22-of-74 5.0 GB: 23-of-74 5.0 GB: 24-of-74 5.0 GB: 25-of-74 5.0 GB: 26-of-74 5.0 GB: 27-of-74 5.0 GB: 28-of-74 5.0 GB: 29-of-74 5.0 GB: 30-of-74 5.0 GB: 31-of-74 5.0 GB: 32-of-74 5.0 GB: 33-of-74 5.0 GB: 34-of-74 5.0 GB: 35-of-74 5.0 GB: 36-of-74 5.0 GB: 37-of-74 5.0 GB: 38-of-74 5.0 GB: 39-of-74 5.0 GB: 40-of-74 5.0 GB: 41-of-74 5.0 GB: 42-of-74 5.0 GB: 43-of-74 5.0 GB: 44-of-74 5.0 GB: 45-of-74
Supported Languages	en zh
AWQ Quantization	Yes
Quantization Type	awq
Model Architecture	DeepseekV3ForCausalLM
License	mit
Context Length	163840
Model Max Length	163840
Transformers Version	4.49.0.dev0
Vocabulary Size	129280
Torch Data Type	bfloat16

Rank the DeepSeek R1 AWQ 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 AWQ

Best Alternatives	Context / RAM	Downloads	Likes
DeepSeek V3 AWQ	160K / 351.9 GB	1809	34
DeepSeek R1 0528 AWQ	160K / 205 GB	8496	15
DeepSeek V3.0324 AWQ	160K / 351.9 GB	4061	22
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 GGUF	160K / GB	27242	1116
DeepSeek R1 0528 FP4	160K / 107 GB	72991	36
DeepSeek R1 0528 GGUF	160K / 162 GB	7195	202

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