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

KULLM3 AWQ by taeminlee

 »  All LLMs  »  taeminlee  »  KULLM3 AWQ   URL Share it on

  7   

KULLM3 AWQ is an open-source language model by taeminlee. Features: 10.7b LLM, VRAM: 6GB, Context: 4K, License: cc-by-nc-4.0, Quantized, LLM Explorer Score: 0.12.

  4-bit   Awq   Base model:nlpai-lab/kullm3 Base model:quantized:nlpai-lab...   Conversational   Deploy:azure   En   Endpoints compatible   Ko   Llama   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/taeminlee/KULLM3-awq 

KULLM3 AWQ Benchmarks

LLME Score: 0.1249
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
KULLM3 AWQ (taeminlee/KULLM3-awq)
🌟 Advertise your project 🚀

KULLM3 AWQ Parameters and Internals

Model Type 
language model
Additional Notes 
The model's quantization was carried out in a custom branch of AutoAWQ and works using vllm. It may not work with other frameworks as they were not tested.
Supported Languages 
Korean (high proficiency), English (high proficiency)
Training Details 
Data Sources:
vicgalle/alpaca-gpt4, Mixed Korean instruction data (gpt-generated, hand-crafted, etc)
Data Volume:
66000+ examples
Input Output 
Performance Tips:
Without system prompt used in the training phase, KULLM would show lower performance than expected.
LLM NameKULLM3 AWQ
Repository 🤗https://huggingface.co/taeminlee/KULLM3-awq 
Base Model(s)  KULLM3   upstage/SOLAR-10.7B-v1.0   nlpai-lab/KULLM3   upstage/SOLAR-10.7B-v1.0
Model Size10.7b
Required VRAM6 GB
Updated2026-05-04
Maintainertaeminlee
Model Typellama
Model Files  5.0 GB: 1-of-2   1.0 GB: 2-of-2
Supported Languagesen ko
AWQ QuantizationYes
Quantization Typeawq
Model ArchitectureLlamaForCausalLM
Licensecc-by-nc-4.0
Context Length4096
Model Max Length4096
Transformers Version4.39.3
Tokenizer ClassLlamaTokenizer
Padding Token</s>
Vocabulary Size32000
Torch Data Typefloat16

Best Alternatives to KULLM3 AWQ

Best Alternatives
Context / RAM
Downloads
Likes
...auerkrautLM SOLAR Instruct AWQ8K / 6 GB142
...auerkrautLM SOLAR Instruct AWQ8K / 6 GB100
STOCK SOLAR 10.7B AWQ4K / 6 GB50
Tess 10.7B V1.5B AWQ4K / 6 GB94
Sensualize Solar 10.7B AWQ4K / 6 GB52
Nous Hermes 2 SOLAR 10.7B AWQ4K / 6 GB1210
CarbonVillain En 10.7B V4 AWQ4K / 6 GB51
SOLAR 10.7B Instruct V1.0 AWQ4K / 6 GB1935
Sakura SOLAR Instruct AWQ4K / 6 GB71
LMCocktail 10.7B V1 AWQ4K / 6 GB51
Note: green Score (e.g. "73.2") means that the model is better than taeminlee/KULLM3-awq.

Rank the KULLM3 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? 53999 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