LLM EXPLORER 59,420 MODELS INDEXED

Nanbeige 16B Base 32K GPTQ by TheBloke

By TheBloke · 17 downloads
HF Repo ↗ Copy URL

Nanbeige 16B Base 32K GPTQ is an open-source language model by TheBloke. Features: 16b LLM, VRAM: 9.2GB, Context: 4K, License: apache-2.0, Quantized, LLM Explorer Score: 0.1.

  4-bit   Custom code   En   Gptq   Nanbeige   Quantized   Region:us   Safetensors   Sharded   Tensorflow   Zh
Model Card on HF 🤗: https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GPTQ 

Nanbeige 16B Base 32K GPTQ Parameters and Internals

Model Type 
nanbeige
Use Cases 
Limitations:
May generate unexpected outputs including harmful content such as bias or discrimination
Supported Languages 
en (English), zh (Chinese)
Training Details 
Data Sources:
high-quality internet corpus, various books, code
Data Volume:
2.5T Tokens
Methodology:
YaRN interpolation for position encoding, full parameter incremental pre-training with 20B tokens
Context Length:
32000
LLM NameNanbeige 16B Base 32K GPTQ
Repository 🤗https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GPTQ 
Model NameNanbeige 16B Base 32K
Model CreatorNanbeige LLM Lab
Base Model(s)  Nanbeige/Nanbeige-16B-Base-32K   Nanbeige/Nanbeige-16B-Base-32K
Model Size16b
Required VRAM9.2 GB
Updated2026-07-28
MaintainerTheBloke
Model Typenanbeige
Model Files  5.0 GB: 1-of-2   4.2 GB: 2-of-2
Supported Languagesen zh
GPTQ QuantizationYes
Quantization Typegptq
Model ArchitectureNanbeigeForCausalLM
Licenseapache-2.0
Context Length4096
Model Max Length4096
Transformers Version4.35.2
Tokenizer ClassNanbeigeTokenizer
Beginning of Sentence Token<s>
End of Sentence Token</s>
Unk Token<unk>
Vocabulary Size59392
Torch Data Typebfloat16

Best Alternatives to Nanbeige 16B Base 32K GPTQ

Best Alternatives
Context / RAM
Downloads
Likes
Nanbeige 16B Base GPTQ4K / 9.2 GB122
Nanbeige 16B Chat GPTQ4K / 9.2 GB61
Nanbeige 16B Chat 32K GPTQ4K / 9.2 GB103