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

Falcon 40B Instruct GPTQ by TheBloke

 »  All LLMs  »  TheBloke  »  Falcon 40B Instruct GPTQ   URL Share it on

  323   
  Arxiv:1911.02150   Arxiv:2005.14165   Arxiv:2104.09864   Arxiv:2205.14135   4-bit   Autotrain compatible   Custom code Dataset:tiiuae/falcon-refinedw...   En   Gptq   Instruct   Quantized   Refinedweb   Region:us   Safetensors
Model Card on HF 🤗: https://huggingface.co/TheBloke/falcon-40b-instruct-GPTQ 

Falcon 40B Instruct GPTQ Benchmarks

LLME Score: 0.10318
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Falcon 40B Instruct GPTQ (TheBloke/falcon-40b-instruct-GPTQ)
🌟 Advertise your project 🚀

Falcon 40B Instruct GPTQ Parameters and Internals

Model Type 
Causal decoder-only, Quantized model
Use Cases 
Areas:
Research, Personal use
Applications:
Chatbots, Customer service
Primary Use Cases:
Ready-to-use chat/instruct
Limitations:
Production use without risk assessment
Considerations:
Is mostly trained in English; does not generalize well to other languages
Supported Languages 
English (high proficiency), French (medium proficiency)
Training Details 
Data Sources:
Baize, RefinedWeb
Data Volume:
150M tokens
Methodology:
Finetuned from Falcon-7B
Context Length:
2048
Hardware Used:
64 A100 40GB GPUs
Model Architecture:
GPT-3 inspired with rotary embeddings, multiquery and FlashAttention
Safety Evaluation 
Risk Categories:
Bias, Stereotypes
Responsible Ai Considerations 
Fairness:
Contains stereotypes and biases from web corpora
Mitigation Strategies:
Develop guardrails and precautions for production use
Input Output 
Input Format:
Prompts using 'A helpful assistant' template
Accepted Modalities:
Text
Output Format:
Generated text responses
Performance Tips:
Very slow, expect around 0.7 tokens/s
Release Notes 
Version:
4bit GPTQ
Notes:
Quantised to 4bit using AutoGPTQ
LLM NameFalcon 40B Instruct GPTQ
Repository 🤗https://huggingface.co/TheBloke/falcon-40b-instruct-GPTQ 
Base Model(s)  Medfalcon 40B Lora   nmitchko/medfalcon-40b-lora
Model Size40b
Required VRAM22.5 GB
Updated2025-08-21
MaintainerTheBloke
Model TypeRefinedWeb
Instruction-BasedYes
Model Files  22.5 GB
Supported Languagesen
GPTQ QuantizationYes
Quantization Typegptq
Model ArchitectureRWForCausalLM
Licenseapache-2.0
Model Max Length2048
Transformers Version4.29.2
Is Biased0
Tokenizer ClassPreTrainedTokenizerFast
Vocabulary Size65024
Torch Data Typebfloat16

Best Alternatives to Falcon 40B Instruct GPTQ

Best Alternatives
Context / RAM
Downloads
Likes
Falcon 40B Instruct GPTQ0K / 22.5 GB141
...truct GPTQ Inference Endpoints0K / 22.5 GB152
Falcon 40B Instruct 8bit0K / 41.8 GB136
...alcon 40B Instruct W4 G128 AWQ0K / 22.3 GB142
H2ogpt Oig Oasst1 Falcon 40B0K / 82.5 GB186

Rank the Falcon 40B Instruct GPTQ 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? 50804 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.
Release v20241124
  Support LLM Explorer