LLM EXPLORER 61,491 MODELS INDEXED

Mellum2 12B A2.5B Thinking Mlx 4bit by jedisct1

By jedisct1 · 521 downloads
HF Repo ↗ Copy URL

Mellum2 12B A2.5B Thinking Mlx 4bit is an open-source language model by jedisct1. Features: 12b LLM, VRAM: 7.3GB, Context: 128K, License: apache-2.0, Quantized, LLM Explorer Score: 0.26.

  4-bit   4bit Base model:jetbrains/mellum2-1... Base model:quantized:jetbrains...   Code   Conversational   En   Mellum   Mlx   Moe   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/jedisct1/Mellum2-12B-A2.5B-Thinking-mlx-4bit 

Mellum2 12B A2.5B Thinking Mlx 4bit Parameters and Internals

LLM NameMellum2 12B A2.5B Thinking Mlx 4bit
Repository 🤗https://huggingface.co/jedisct1/Mellum2-12B-A2.5B-Thinking-mlx-4bit 
Base Model(s)  JetBrains/Mellum2-12B-A2.5B-Thinking   JetBrains/Mellum2-12B-A2.5B-Thinking
Model Size12b
Required VRAM7.3 GB
Updated2026-08-08
Maintainerjedisct1
Model Typemellum
Model Files  5.3 GB: 1-of-2   2.0 GB: 2-of-2
Supported Languagesen
Quantization Type4bit
Model ArchitectureMellumForCausalLM
Licenseapache-2.0
Context Length131072
Model Max Length131072
Transformers Version5.8.1
Tokenizer ClassTokenizersBackend
Padding Token<|endoftext|>
Vocabulary Size98304

Best Alternatives to Mellum2 12B A2.5B Thinking Mlx 4bit

Best Alternatives
Context / RAM
Downloads
Likes
...ellum2 12B A2.5B Thinking 4bit128K / 6.8 GB1261
...ellum2 12B A2.5B Instruct 4bit128K / 6.8 GB141
Mellum2 12B A2.5B Thinking128K / 24.3 GB6221327
Mellum2 12B A2.5B Base128K / 24.3 GB1675323
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
Note: green Score (e.g. "73.2") means that the model is better than jedisct1/Mellum2-12B-A2.5B-Thinking-mlx-4bit.