What are the hardware requirements for MedraN E4B?

MedraN E4B requires approximately 15.8 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed MedraN E4B and how large is it?

MedraN E4B is developed by nicoboss, a model with 7.8b 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 MedraN E4B?

MedraN E4B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

MedraN E4B by nicoboss — VRAM 15.8GB

Name: MedraN E4B
Author: nicoboss

MedraN E4B is an open-source language model by nicoboss. Features: 7.8b LLM, VRAM: 15.8GB, LLM Explorer Score: 0.19.

Gemma3n Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/nicoboss/MedraN-E4B

MedraN E4B Parameters and Internals

LLM Name	MedraN E4B
Repository 🤗	https://huggingface.co/nicoboss/MedraN-E4B
Model Size	7.8b
Required VRAM	15.8 GB
Updated	2026-04-22
Maintainer	nicoboss
Model Type	gemma3n
Model Files	5.0 GB: 1-of-4 4.6 GB: 2-of-4 5.0 GB: 3-of-4 1.2 GB: 4-of-4
Model Architecture	Gemma3nForConditionalGeneration
Transformers Version	4.53.2
Tokenizer Class	GemmaTokenizer
Padding Token	<pad>
Torch Data Type	bfloat16

Rank the MedraN E4B 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 MedraN E4B

Best Alternatives	Context / RAM	Downloads	Likes
Gemma 3n E4B It	0K / 15.8 GB	35121	919
...mma 3n E4B It Unsloth Bnb 4bit	0K / 9.8 GB	1649	9
Gemma 3n E4B Unsloth Bnb 4bit	0K / 9.8 GB	189	4

Note: green Score (e.g. "73.2") means that the model is better than nicoboss/MedraN-E4B.