What are the hardware requirements for MiniCPM5 2B Heretic Abliterated?

MiniCPM5 2B Heretic Abliterated requires approximately 5 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 MiniCPM5 2B Heretic Abliterated and how large is it?

MiniCPM5 2B Heretic Abliterated is developed by insraq, a model with 2b 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 MiniCPM5 2B Heretic Abliterated?

MiniCPM5 2B Heretic Abliterated is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

MiniCPM5 2B Heretic Abliterated by insraq — VRAM 5GB, 128K context

Name: MiniCPM5 2B Heretic Abliterated
Author: insraq

MiniCPM5 2B Heretic Abliterated is an open-source language model by insraq. Features: 2b LLM, VRAM: 5GB, Context: 128K, License: apache-2.0, LLM Explorer Score: 0.28.

Arxiv:2506.07900 Arxiv:2602.09003 Abliterated Conversational Dataset:openbmb/ultra-fineweb Dataset:openbmb/ultra-fineweb-... Dataset:openbmb/ultradata-code Dataset:openbmb/ultradata-math Dataset:openbmb/ultradata-rl-2... Dataset:openbmb/ultradata-sft-... Dataset:openbmb/ultradata-sft-... Dataset:openbmb/ultrax-preview Decensored Edge-ai En Endpoints compatible Heretic Llama Long-context Minicpm Minicpm5 On-device Region:us Reproducible Safetensors Sharded Tensorflow Tool-calling Uncensored Zh

Model Card on HF 🤗: https://huggingface.co/insraq/MiniCPM5-2B-heretic-abliterated

MiniCPM5 2B Heretic Abliterated Parameters and Internals

LLM Name	MiniCPM5 2B Heretic Abliterated
Repository 🤗	https://huggingface.co/insraq/MiniCPM5-2B-heretic-abliterated
Model Size	2b
Required VRAM	5 GB
Updated	2026-09-09
Maintainer	insraq
Model Type	llama
Model Files	5.0 GB: 1-of-2 0.0 GB: 2-of-2
Supported Languages	en zh
Model Architecture	LlamaForCausalLM
License	apache-2.0
Context Length	131072
Model Max Length	131072
Transformers Version	5.14.1
Tokenizer Class	TokenizersBackend
Padding Token	</s>
Vocabulary Size	130560

Rank the MiniCPM5 2B Heretic Abliterated 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 MiniCPM5 2B Heretic Abliterated

Best Alternatives	Context / RAM	Downloads	Likes
MiniCPM5 2B Base	512K / 5 GB	54	17
MiniCPM5 2B	128K / 5 GB	2879	623
Huihui MiniCPM5 2B Abliterated	128K / 5 GB	1609	17
Xor Lite	128K / 5 GB	484	3
...iniCPM5 2B RotSVDMix Quantized	128K / 2 GB	1389	1
MiniCPM5 2B MLX	128K / 1.4 GB	288	24
BananaMind CodeQ 1.3 2B	128K / 5.1 GB	718	3
Pjev 2B	128K / 5 GB	356	1
Sharp MiniCPM5 2B MLX OQ6e	128K / 2.1 GB	240	2
...iniCPM5 2B SFT Pashto Instruct	128K / 5 GB	1076	0

Note: green Score (e.g. "73.2") means that the model is better than insraq/MiniCPM5-2B-heretic-abliterated.