LLM EXPLORER 58,608 MODELS INDEXED

LongCat Flash Lite 8bit by mlx-community

By mlx-community · 57 downloads
HF Repo ↗ Copy URL

LongCat Flash Lite 8bit is an open-source language model by mlx-community. Features: 68.6b LLM, VRAM: 73.3GB, Context: 320K, License: mit, Quantized, LLM Explorer Score: 0.22.

  8-bit   8bit Base model:meituan-longcat/lon... Base model:quantized:meituan-l...   Conversational   Custom code   Longcat flash ngram   Mlx   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/mlx-community/LongCat-Flash-Lite-8bit 

LongCat Flash Lite 8bit Parameters and Internals

LLM NameLongCat Flash Lite 8bit
Repository 🤗https://huggingface.co/mlx-community/LongCat-Flash-Lite-8bit 
Base Model(s)  LongCat Flash Lite   meituan-longcat/LongCat-Flash-Lite
Model Size68.6b
Required VRAM73.3 GB
Updated2026-08-10
Maintainermlx-community
Model Typelongcat_flash_ngram
Model Files  3.2 GB: 1-of-20   2.8 GB: 2-of-20   2.8 GB: 3-of-20   2.8 GB: 4-of-20   2.8 GB: 5-of-20   2.8 GB: 6-of-20   2.8 GB: 7-of-20   2.8 GB: 8-of-20   2.8 GB: 9-of-20   2.8 GB: 10-of-20   2.8 GB: 11-of-20   5.4 GB: 12-of-20   4.7 GB: 13-of-20   4.7 GB: 14-of-20   4.7 GB: 15-of-20   5.4 GB: 16-of-20   4.6 GB: 17-of-20   4.7 GB: 18-of-20   4.7 GB: 19-of-20   3.2 GB: 20-of-20
Quantization Type8bit
Model ArchitectureLongcatFlashNgramForCausalLM
Licensemit
Context Length327680
Model Max Length327680
Transformers Version4.57.6
Tokenizer ClassTokenizersBackend
Padding Token<longcat_pad>
Vocabulary Size131072
Torch Data Typebfloat16

Best Alternatives to LongCat Flash Lite 8bit

Best Alternatives
Context / RAM
Downloads
Likes
LongCat Flash Lite 6bit320K / 55.5 GB822
LongCat Flash Lite 4bit DWQ320K / 38.5 GB251
...d Heretic Native MTP Preserved320K / 136.8 GB02
Note: green Score (e.g. "73.2") means that the model is better than mlx-community/LongCat-Flash-Lite-8bit.