LLM EXPLORER 58,133 MODELS INDEXED

Ling 3.0 Flash W4A4 NVFP4 by sakamakismile

By sakamakismile · 341 downloads
HF Repo ↗ Copy URL

Ling 3.0 Flash W4A4 NVFP4 is an open-source language model by sakamakismile. Features: 74.6b LLM, VRAM: 81.3GB, Context: 128K, License: mit, LLM Explorer Score: 0.32.

  8-bit   Bailing hybrid Base model:inclusionai/ling-3.... Base model:quantized:inclusion...   Compressed-tensors   Conversational   Custom code   Hybrid-linear-attention   Moe   Mtp   Nvfp4   Region:us   Safetensors   Sharded   Tensorflow   Vllm   W4a4
Model Card on HF 🤗: https://huggingface.co/sakamakismile/Ling-3.0-flash-W4A4-NVFP4 

Ling 3.0 Flash W4A4 NVFP4 Parameters and Internals

LLM NameLing 3.0 Flash W4A4 NVFP4
Repository 🤗https://huggingface.co/sakamakismile/Ling-3.0-flash-W4A4-NVFP4 
Base Model(s)  inclusionAI/Ling-3.0-flash   inclusionAI/Ling-3.0-flash
Model Size74.6b
Required VRAM81.3 GB
Updated2026-08-11
Maintainersakamakismile
Model Typebailing_hybrid
Model Files  10.0 GB: 1-of-9   10.0 GB: 2-of-9   10.0 GB: 3-of-9   10.0 GB: 4-of-9   9.2 GB: 5-of-9   10.0 GB: 6-of-9   10.0 GB: 7-of-9   10.0 GB: 8-of-9   2.1 GB: 9-of-9
Model ArchitectureBailingMoeV3ForCausalLM
Licensemit
Context Length131072
Model Max Length131072
Transformers Version5.14.1+solaropen2
Tokenizer ClassTokenizersBackend
Padding Token<|endoftext|>
Vocabulary Size157184