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

Karasu Mixtral 8x22B V0.1 AWQ by lightblue

 »  All LLMs  »  lightblue  »  Karasu Mixtral 8x22B V0.1 AWQ   URL Share it on

  4   

Karasu Mixtral 8x22B V0.1 AWQ is an open-source language model by lightblue. Features: 140.6b LLM, VRAM: 73.7GB, Context: 64K, MoE, Quantized, Merged, LLM Explorer Score: 0.13.

  Merged Model   4-bit   Awq   Conversational   Endpoints compatible   Mixtral   Moe   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/lightblue/Karasu-Mixtral-8x22B-v0.1-AWQ 

Karasu Mixtral 8x22B V0.1 AWQ Benchmarks

LLME Score: 0.12954
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Karasu Mixtral 8x22B V0.1 AWQ (lightblue/Karasu-Mixtral-8x22B-v0.1-AWQ)
🌟 Advertise your project 🚀

Karasu Mixtral 8x22B V0.1 AWQ Parameters and Internals

Model Type 
Causal Language Model
Additional Notes 
Quantized using AutoAWQForCausalLM with 4-bit configuration, zero_point enabled, q_group_size of 128, and version GEMM.
LLM NameKarasu Mixtral 8x22B V0.1 AWQ
Repository 🤗https://huggingface.co/lightblue/Karasu-Mixtral-8x22B-v0.1-AWQ 
Merged ModelYes
Model Size140.6b
Required VRAM73.7 GB
Updated2026-03-30
Maintainerlightblue
Model Typemixtral
Model Files  5.0 GB: 1-of-15   5.0 GB: 2-of-15   5.0 GB: 3-of-15   5.0 GB: 4-of-15   5.0 GB: 5-of-15   5.0 GB: 6-of-15   5.0 GB: 7-of-15   5.0 GB: 8-of-15   5.0 GB: 9-of-15   5.0 GB: 10-of-15   5.0 GB: 11-of-15   5.0 GB: 12-of-15   5.0 GB: 13-of-15   5.0 GB: 14-of-15   3.7 GB: 15-of-15
AWQ QuantizationYes
Quantization Typeawq
Model ArchitectureMixtralForCausalLM
Context Length65536
Model Max Length65536
Transformers Version4.38.2
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size32000
Torch Data Typefloat16

Best Alternatives to Karasu Mixtral 8x22B V0.1 AWQ

Best Alternatives
Context / RAM
Downloads
Likes
...ixtral 8x22B Instruct V0.1 AWQ64K / 73.7 GB189113
WizardLM 2 8x22B AWQ64K / 73.7 GB913
Mixtral 8x22B Instruct V0.164K / 221.4 GB34298748
Zephyr Orpo 141B A35b V0.164K / 207.2 GB91269
WizardLM 2 8x22B64K / 216.8 GB9728411
Mixtral 8x22B V0.164K / 221.6 GB6061236
Mixtral 8x22B V0.164K / 212 GB237672
Mixtral 8x22B V0.364K / 221.4 GB623
XLAM 8x22b R64K / 211.8 GB908845
...ixtral 8x22B Instruct V0.1 FP864K / 140.9 GB5480
Note: green Score (e.g. "73.2") means that the model is better than lightblue/Karasu-Mixtral-8x22B-v0.1-AWQ.

Rank the Karasu Mixtral 8x22B V0.1 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? 52628 in total.

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