What are the hardware requirements for MXLewd L2 20B?

MXLewd L2 20B requires approximately 40.7 GB of VRAM and supports a context window of 4K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed MXLewd L2 20B and how large is it?

MXLewd L2 20B is developed by Undi95, a model with 20b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does MXLewd L2 20B perform on standard benchmarks?

MXLewd L2 20B has the following published scores: MMLU 57.36. Compare against reference models on this page or on the LLM Explorer leaderboards.

MXLewd L2 20B by Undi95 — VRAM 40.7GB, 4K context

Name: MXLewd L2 20B
Author: Undi95

MXLewd L2 20B is an open-source language model by Undi95. Features: 20b LLM, VRAM: 40.7GB, Context: 4K, License: cc-by-nc-4.0, LLM Explorer Score: 0.11, Arc: 63.2, HellaSwag: 85.3, MMLU: 57.4, GSM8K: 10.9.

Endpoints compatible Llama Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Undi95/MXLewd-L2-20B

MXLewd L2 20B Parameters and Internals

Model Type

Text Generation

Additional Notes

This model consists of a merge of several other models, utilizing different ratios and parts. It is benchmarked on several datasets with varying shot numbers.

Supported Languages

English (Proficient)

Input Output

Input Format:

Alpaca Prompt

Accepted Modalities:

text

Output Format:

text

LLM Name	MXLewd L2 20B
Repository 🤗	https://huggingface.co/Undi95/MXLewd-L2-20B
Model Size	20b
Required VRAM	40.7 GB
Updated	2026-08-08
Maintainer	Undi95
Model Type	llama
Model Files	10.3 GB: 1-of-5 10.0 GB: 2-of-5 10.0 GB: 3-of-5 9.7 GB: 4-of-5 0.7 GB: 5-of-5
Model Architecture	LlamaForCausalLM
License	cc-by-nc-4.0
Context Length	4096
Model Max Length	4096
Transformers Version	4.33.2
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000
Torch Data Type	bfloat16

Rank the MXLewd L2 20B 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

Quantized Models of the MXLewd L2 20B

Model	Likes	Downloads	VRAM
MXLewd L2 20B GGUF	30	544	8 GB
MXLewd L2 20B GPTQ	9	8	10 GB
MXLewd L2 20B AWQ	3	9	10 GB

Best Alternatives to MXLewd L2 20B

Best Alternatives	Context / RAM	Downloads	Likes
Internlm2 5 20B Llamafied	256K / 39.9 GB	7	5
EAGLE3 GPT Oss 20B	128K / 0.7 GB	507	0
Stellaris Internlm2 20B R512	32K / 39.8 GB	12	3
Internlm2 Chat 20B Llama Old	32K / 39.6 GB	8	3
Internlm2 Base 20B Llama	32K / 39.6 GB	6	3
Internlm2 Base 20B Llama	32K / 39.6 GB	6	0
Deita 20B	32K / 39.8 GB	63	1
Internlm2 20B Llama	32K / 39.6 GB	81	20
Bagel 20B V04 Llama	32K / 39.6 GB	6	7
Bagel DPO 20B V04 Llama	32K / 39.6 GB	7	3

Note: green Score (e.g. "73.2") means that the model is better than Undi95/MXLewd-L2-20B.