LLM EXPLORER 64,561 MODELS INDEXED

Mellum2.1 12B A2.5B Thinking MXFP4 by FP8Enjoyer

By FP8Enjoyer · 47 downloads
HF Repo ↗ Copy URL

Mellum2.1 12B A2.5B Thinking MXFP4 is an open-source language model by FP8Enjoyer. Features: 12b LLM, VRAM: 7.1GB, Context: 128K, License: apache-2.0.

  8-bit Base model:jetbrains/mellum2.1... Base model:quantized:jetbrains...   Blackwell   Compressed-tensors   Conversational   En   Mellum   Moe   Mxfp4   Pytorch   Quantized   Region:us   Safetensors   Windows
Model Card on HF 🤗: https://huggingface.co/FP8Enjoyer/Mellum2.1-12B-A2.5B-Thinking-MXFP4 

Mellum2.1 12B A2.5B Thinking MXFP4 Parameters and Internals

LLM NameMellum2.1 12B A2.5B Thinking MXFP4
Repository 🤗https://huggingface.co/FP8Enjoyer/Mellum2.1-12B-A2.5B-Thinking-MXFP4 
Base Model(s)  JetBrains/Mellum2.1-12B-A2.5B-Thinking   JetBrains/Mellum2.1-12B-A2.5B-Thinking
Model Size12b
Required VRAM7.1 GB
Updated2026-10-17
MaintainerFP8Enjoyer
Model Typemellum
Model Files  7.1 GB
Supported Languagesen
Model ArchitectureMellumForCausalLM
Licenseapache-2.0
Context Length131072
Model Max Length131072
Transformers Version5.8.1
Tokenizer ClassPreTrainedTokenizerFast
Padding Token<|endoftext|>
Vocabulary Size98304

Best Alternatives to Mellum2.1 12B A2.5B Thinking MXFP4

Best Alternatives
Context / RAM
Downloads
Likes
Mellum2.1 12B A2.5B Thinking128K / 24.3 GB80399
Mellum2 12B A2.5B Thinking128K / 24.3 GB6221327
Mellum2 12B A2.5B Base128K / 24.3 GB1675323
...1 12B A2.5B Thinking Mlx Nvfp4128K / 6.8 GB1951
Mellum2 12B A2.5B Instruct128K / 24.3 GB449882
...llum2 12B A2.5B Thinking Mxfp4128K / 6.4 GB3011
...llum2 12B A2.5B Instruct Mxfp4128K / 6.4 GB1661
...12B A2.5B Thinking Abliterated128K / 24.3 GB765
...lum2.1 12B A2.5B Thinking OQ5e128K / 8.6 GB6492
...lum2.1 12B A2.5B Thinking OQ8e128K / 12.8 GB2551
Note: green Score (e.g. "73.2") means that the model is better than FP8Enjoyer/Mellum2.1-12B-A2.5B-Thinking-MXFP4.