What are the hardware requirements for Pygmalion 2 13B?

Pygmalion 2 13B requires approximately 26 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 Pygmalion 2 13B and how large is it?

Pygmalion 2 13B is developed by PygmalionAI, a model with 13b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Pygmalion 2 13B perform on standard benchmarks?

Pygmalion 2 13B has the following published scores: MMLU 56.02. Compare against reference models on this page or on the LLM Explorer leaderboards.

Pygmalion 2 13B by PygmalionAI — VRAM 26GB, 4K context

Name: Pygmalion 2 13B
Author: PygmalionAI

Pygmalion 2 13B is an open-source language model by PygmalionAI. Features: 13b LLM, VRAM: 26GB, Context: 4K, License: llama2, Instruction-Based, HF Score: 55.1, LLM Explorer Score: 0.13, Arc: 60.3, HellaSwag: 82.4, MMLU: 56, TruthfulQA: 42.2, WinoGrande: 78.1, GSM8K: 11.8.

Dataset:databricks/databricks-... Dataset:jondurbin/airoboros-gp... Dataset:norquinal/claude multi... Dataset:open-orca/openorca Dataset:pygmalionai/pippa Deploy:azure En Instruct Llama Pytorch Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/PygmalionAI/pygmalion-2-13b

Pygmalion 2 13B Benchmarks

ARC: 60.32 vs 96.7 (so35)^-37.6%

HellaSwag: 82.37 vs 95.3 (gpt4)^-13.6%

MMLU: 56.02 vs 88.3 (so35)^-36.6%

TruthfulQA: 42.22 vs 59 (gpt4)^-28.4%

WinoGrande: 78.06 vs 87.5 (gpt4)^-10.8%

GSM8K: 11.75 vs 96.4 (so35)^-87.8%

LLME Score: 0.13106

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

Pygmalion 2 13B (PygmalionAI/pygmalion-2-13b)

🌟 Advertise your project 🚀

Pygmalion 2 13B Parameters and Internals

Model Type

text generation, instruct

Use Cases

Areas:

fictional writing for entertainment purposes

Limitations:

Not fine-tuned to be safe and harmless, May produce socially unacceptable or undesirable text, Outputs might often be factually wrong or misleading

Additional Notes

The model is biased towards fiction writing and conversation. It was known as Metharme and aimed for roles involving conversation, roleplaying, and storywriting that can be guided using natural language.

Supported Languages

English (High)

Training Details

Data Sources:

PygmalionAI/PIPPA, Open-Orca/OpenOrca, Norquinal/claude_multiround_chat_30k, jondurbin/airoboros-gpt4-1.4.1, databricks/databricks-dolly-15k

Methodology:

Supervised fine-tuning over a mixture of regular instruction data alongside roleplay, fictional stories and conversations with synthetically generated instructions attached.

Input Output

Input Format:

<|system|>, <|user|>, <|model|> prompts

Accepted Modalities:

text

LLM Name	Pygmalion 2 13B
Repository 🤗	https://huggingface.co/PygmalionAI/pygmalion-2-13b
Model Size	13b
Required VRAM	26 GB
Updated	2026-04-12
Maintainer	PygmalionAI
Model Type	llama
Instruction-Based	Yes
Model Files	9.9 GB: 1-of-3 9.9 GB: 2-of-3 6.2 GB: 3-of-3 9.9 GB: 1-of-3 9.9 GB: 2-of-3 6.2 GB: 3-of-3
Supported Languages	en
Model Architecture	LlamaForCausalLM
License	llama2
Context Length	4096
Model Max Length	4096
Transformers Version	4.33.0.dev0
Tokenizer Class	LlamaTokenizer
Beginning of Sentence Token	<s>
End of Sentence Token	</s>
Unk Token	<unk>
Vocabulary Size	32000
Torch Data Type	bfloat16

Quantized Models of the Pygmalion 2 13B

Model	Likes	Downloads	VRAM
Pygmalion 2 13B GGUF	32	1440	5 GB
Pygmalion 2 13B AWQ	6	1104	7 GB
Pygmalion 2 13B GPTQ	41	12	7 GB

Best Alternatives to Pygmalion 2 13B

Best Alternatives	Context / RAM	Downloads	Likes
NexusRaven V2 13B	16K / 26 GB	80	470
CodeLlama 13B Instruct Hf	16K / 26 GB	5282	159
CodeLlama 13B MORepair	16K / 26 GB	12	2
CodeLlama 13B Instruct Hf	16K / 26 GB	234	27
TableLLM 13B	16K / 26 GB	306	32
NexusRaven 13B	16K / 26 GB	17	105
Panda Coder 13B	16K / 26 GB	20	14
... Llama 2 13B Instruct Text2sql	16K / 26 GB	42	28
Gen Sim	16K / 0.3 GB	8	3
Llama 3 13B Instruct Ft	8K / 26.1 GB	10	2

Note: green Score (e.g. "73.2") means that the model is better than PygmalionAI/pygmalion-2-13b.

Rank the Pygmalion 2 13B Capabilities

🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟

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

What open-source LLMs or SLMs are you in search of? 53205 in total.

Email us: info@extractum.io. Our Privacy Policy | Terms and Conditions | Suggest an improvement.

Our Social Media →

Original data from HuggingFace, OpenCompass and various public git repos.

Check out Ag3ntum — our secure, self-hosted AI agent for server management.

Release v20260328a

Support LLM Explorer

Pygmalion 2 13B by PygmalionAI