What are the hardware requirements for Mellum2.1 12B A2.5B Thinking OQ5e?

Mellum2.1 12B A2.5B Thinking OQ5e requires approximately 8.6 GB of VRAM and supports a context window of 128K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Mellum2.1 12B A2.5B Thinking OQ5e and how large is it?

Mellum2.1 12B A2.5B Thinking OQ5e is developed by niteg, a model with 12b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Mellum2.1 12B A2.5B Thinking OQ5e?

Mellum2.1 12B A2.5B Thinking OQ5e is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Mellum2.1 12B A2.5B Thinking OQ5e by niteg — VRAM 8.6GB, 128K context

Name: Mellum2.1 12B A2.5B Thinking OQ5e
Author: niteg

Mellum2.1 12B A2.5B Thinking OQ5e is an open-source language model by niteg. Features: 12b LLM, VRAM: 8.6GB, Context: 128K, License: apache-2.0.

Arxiv:2605.31268 5-bit Base model:jetbrains/mellum2.1... Base model:quantized:jetbrains... Coding Conversational En Mellum Mlx Moe Oq Quantized Region:us Safetensors Sharded Tensorflow Thinking

Model Card on HF 🤗: https://huggingface.co/niteg/Mellum2.1-12B-A2.5B-Thinking-oQ5e

Mellum2.1 12B A2.5B Thinking OQ5e Parameters and Internals

LLM Name	Mellum2.1 12B A2.5B Thinking OQ5e
Repository 🤗	https://huggingface.co/niteg/Mellum2.1-12B-A2.5B-Thinking-oQ5e
Base Model(s)	JetBrains/Mellum2.1-12B-A2.5B-Thinking JetBrains/Mellum2.1-12B-A2.5B-Thinking
Model Size	12b
Required VRAM	8.6 GB
Updated	2026-10-17
Maintainer	niteg
Model Type	mellum
Model Files	5.0 GB: 1-of-2 3.6 GB: 2-of-2
Supported Languages	en
Model Architecture	MellumForCausalLM
License	apache-2.0
Context Length	131072
Model Max Length	131072
Transformers Version	5.8.1
Tokenizer Class	PreTrainedTokenizerFast
Padding Token	<\|endoftext\|>
Vocabulary Size	98304

Rank the Mellum2.1 12B A2.5B Thinking OQ5e capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Mellum2.1 12B A2.5B Thinking OQ5e

Best Alternatives	Context / RAM	Downloads	Likes
Mellum2.1 12B A2.5B Thinking	128K / 24.3 GB	803	99
Mellum2 12B A2.5B Thinking	128K / 24.3 GB	6221	327
Mellum2 12B A2.5B Base	128K / 24.3 GB	16753	23
...1 12B A2.5B Thinking Mlx Nvfp4	128K / 6.8 GB	195	1
Mellum2 12B A2.5B Instruct	128K / 24.3 GB	4498	82
...llum2 12B A2.5B Thinking Mxfp4	128K / 6.4 GB	301	1
...llum2 12B A2.5B Instruct Mxfp4	128K / 6.4 GB	166	1
...12B A2.5B Thinking Abliterated	128K / 24.3 GB	76	5
...um2.1 12B A2.5B Thinking MXFP4	128K / 7.1 GB	47	1
...lum2.1 12B A2.5B Thinking OQ8e	128K / 12.8 GB	255	1

Note: green Score (e.g. "73.2") means that the model is better than niteg/Mellum2.1-12B-A2.5B-Thinking-oQ5e.