LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Leaderboards   Agents   Benchmarks    LLM Hosting    Advertise

DeepSeek V3 AWQ by cognitivecomputations

 »  All LLMs  »  cognitivecomputations  »  DeepSeek V3 AWQ   URL Share it on

  232   

DeepSeek V3 AWQ is an open-source language model by cognitivecomputations. Features: 671b LLM, VRAM: 351.9GB, Context: 160K, License: mit, Quantized, LLM Explorer Score: 0.3, ELO: 1334.

  4-bit   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/QuixiAI/DeepSeek-V3-AWQ 

DeepSeek V3 AWQ Benchmarks

LMSys ELO: 1334 vs 1272 (so35)4.9%
LLME Score: 0.30118
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
DeepSeek V3 AWQ (QuixiAI/DeepSeek-V3-AWQ)

DeepSeek V3 AWQ Parameters and Internals

LLM NameDeepSeek V3 AWQ
Repository 🤗https://huggingface.co/QuixiAI/DeepSeek-V3-AWQ 
Base Model(s)  DeepSeek V3   deepseek-ai/DeepSeek-V3
Model Size671b
Required VRAM351.9 GB
Updated2026-07-25
Maintainercognitivecomputations
Model Typedeepseek_v3
Model Files  10.0 GB: 1-of-36   10.0 GB: 2-of-36   10.0 GB: 3-of-36   10.0 GB: 4-of-36   10.0 GB: 5-of-36   10.0 GB: 6-of-36   10.0 GB: 7-of-36   10.0 GB: 8-of-36   10.0 GB: 9-of-36   10.0 GB: 10-of-36   10.0 GB: 11-of-36   10.0 GB: 12-of-36   10.0 GB: 13-of-36   10.0 GB: 14-of-36   10.0 GB: 15-of-36   10.0 GB: 16-of-36   10.0 GB: 17-of-36   10.0 GB: 18-of-36   10.0 GB: 19-of-36   10.0 GB: 20-of-36   10.0 GB: 21-of-36   10.0 GB: 22-of-36   10.0 GB: 23-of-36   10.0 GB: 24-of-36   10.0 GB: 25-of-36   10.0 GB: 26-of-36   10.0 GB: 27-of-36   10.0 GB: 28-of-36   10.0 GB: 29-of-36   10.0 GB: 30-of-36   10.0 GB: 31-of-36   10.0 GB: 32-of-36   10.0 GB: 33-of-36   10.0 GB: 34-of-36   10.0 GB: 35-of-36   1.9 GB: 36-of-36
Supported Languagesen zh
AWQ QuantizationYes
Quantization Typeawq
Model ArchitectureDeepseekV3ForCausalLM
Licensemit
Context Length163840
Model Max Length163840
Transformers Version4.48.0.dev0
Tokenizer ClassLlamaTokenizer
Padding Token<｜end▁of▁sentence｜>
Vocabulary Size129280
Torch Data Typefloat16

Best Alternatives to DeepSeek V3 AWQ

Best Alternatives
Context / RAM
Downloads
Likes
DeepSeek R1 0528 AWQ160K / 205 GB9919
DeepSeek R1 AWQ160K / 225 GB94187
DeepSeek V3.0324 AWQ160K / 351.9 GB3023
DeepSeek V3.1 Terminus 4bit160K / 190.1 GB452
DeepSeek V3.1 4bit160K / 190.1 GB48406
RakutenAI 3.0 MLX 4bit160K / 198.7 GB282
DeepSeek V3.1 8bit160K / 177.8 GB42683
DeepSeek V3.1 Base 4bit160K / 190.1 GB553
DeepSeek R1 0528 5bit160K / 231 GB663
R1 1776160K / 243.4 GB8552324
Note: green Score (e.g. "73.2") means that the model is better than QuixiAI/DeepSeek-V3-AWQ.

Rank the DeepSeek V3 AWQ Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

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  

What open-source LLMs or SLMs are you in search of? 55449 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, Arena and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a
  Support LLM Explorer