LLM EXPLORER 62,007 MODELS INDEXED

Gemma 4 31B It Qat NVFP4 Mse Blackwell by melcheikh

By melcheikh · 440 downloads
HF Repo ↗ Copy URL

Gemma 4 31B It Qat NVFP4 Mse Blackwell is an open-source language model by melcheikh. Features: 31b LLM, VRAM: 20.5GB, License: gemma, Quantized, LLM Explorer Score: 0.27.

  8-bit Base model:google/gemma-4-31b-... Base model:quantized:google/ge...   Blackwell   Conversational   Fp4   Gemma4   Modelopt   Mse   Nvfp4   Nvidia   Q4   Q4 0   Qat   Quantized   Region:us   Safetensors   Sharded   Tensorflow   Vllm
Model Card on HF 🤗: https://huggingface.co/melcheikh/gemma-4-31B-it-qat-NVFP4-mse-Blackwell 

Gemma 4 31B It Qat NVFP4 Mse Blackwell Parameters and Internals

LLM NameGemma 4 31B It Qat NVFP4 Mse Blackwell
Repository 🤗https://huggingface.co/melcheikh/gemma-4-31B-it-qat-NVFP4-mse-Blackwell 
Base Model(s)  ... 4 31B It Qat Q4 0 Unquantized   google/gemma-4-31B-it-qat-q4_0-unquantized
Model Size31b
Required VRAM20.5 GB
Updated2026-08-29
Maintainermelcheikh
Model Typegemma4
Model Files  10.0 GB: 1-of-3   10.0 GB: 2-of-3   0.5 GB: 3-of-3
Quantization Typeq4|q4_0
Model ArchitectureGemma4ForConditionalGeneration
Licensegemma
Transformers Version5.10.0.dev0
Tokenizer ClassGemmaTokenizer
Padding Token<eos>

Best Alternatives to Gemma 4 31B It Qat NVFP4 Mse Blackwell

Best Alternatives
Context / RAM
Downloads
Likes
Gemma 4 31B It 4bit0K / 18.5 GB4506746
... 4 31B It Qat Q4 0 Unquantized0K / 62.6 GB997239
...emma 4 31B It Unsloth Bnb 4bit0K / 19.2 GB1930625
... 4 31B It Qat Q4 0 Unquantized0K / 62.6 GB20744
Gemma 4 31B It 8bit0K / 33.8 GB321417
Gemma 4 31B 8bit0K / 33.8 GB95724
Gemma 4 31B It UD MLX 4bit0K / 23.2 GB79228
Gemma 4 31B It MLX 8bit0K / 33.8 GB97215
Gemma 4 31B It 6bit0K / 26 GB9067
Gemma 4 31B 4bit0K / 18.5 GB44811
Note: green Score (e.g. "73.2") means that the model is better than melcheikh/gemma-4-31B-it-qat-NVFP4-mse-Blackwell.