What are the hardware requirements for Tofu 3B?

Tofu 3B requires approximately 5.6 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 Tofu 3B and how large is it?

Tofu 3B is developed by jeiku, a model with 3b 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 Tofu 3B?

Tofu 3B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Tofu 3B by jeiku — VRAM 5.6GB, 4K context

Name: Tofu 3B
Author: jeiku

Tofu 3B is an open-source language model by jeiku. Features: 3b LLM, VRAM: 5.6GB, Context: 4K, Merged, LLM Explorer Score: 0.11.

Merged Model Arxiv:2212.04089 Base model:jeiku/alpaca 128 st... Base model:jeiku/bluemoon clea... Base model:jeiku/everything v3... Base model:jeiku/gnosis 256 st... Base model:jeiku/limarp stable... Base model:jeiku/no robots alp... Base model:jeiku/pippa 128 sta... Base model:jeiku/rosa v1 3b Base model:jeiku/rpgpt stablel... Base model:jeiku/theory of min... Base model:jeiku/theory of min... Base model:jeiku/toxic dpo sta... Conversational Custom code Region:us Safetensors Sharded Stablelm epoch Tensorflow

Model Card on HF 🤗: https://huggingface.co/jeiku/Tofu_3B

Tofu 3B Parameters and Internals

Model Type

language model

Additional Notes

This model is a merge of various pre-trained language models using task arithmetic methods. The configuration used includes weights for different model combinations.

Training Details

Methodology:

task_arithmetic merge

LLM Name	Tofu 3B
Repository 🤗	https://huggingface.co/jeiku/Tofu_3B
Base Model(s)	Rosa V1 3B Theory Of Mind 128 StableLM Rosa V1 3B Rosa V1 3B jeiku/PIPPA_128_StableLM Rosa V1 3B jeiku/LimaRP_StableLM Rosa V1 3B Theory Of Mind RP 128 StableLM Rosa V1 3B No Robots Alpaca StableLM Rosa V1 3B jeiku/Alpaca_128_StableLM Rosa V1 3B jeiku/Everything_v3_128_StableLM Rosa V1 3B jeiku/RPGPT_StableLM Rosa V1 3B Toxic DPO StableLM Rosa V1 3B jeiku/Gnosis_256_StableLM Rosa V1 3B Bluemoon Cleaned StableLM jeiku/Rosa_v1_3B jeiku/Theory_of_Mind_128_StableLM jeiku/Rosa_v1_3B jeiku/Rosa_v1_3B jeiku/PIPPA_128_StableLM jeiku/Rosa_v1_3B jeiku/LimaRP_StableLM jeiku/Rosa_v1_3B jeiku/Theory_of_Mind_RP_128_StableLM jeiku/Rosa_v1_3B jeiku/No_Robots_Alpaca_StableLM jeiku/Rosa_v1_3B jeiku/Alpaca_128_StableLM jeiku/Rosa_v1_3B jeiku/Everything_v3_128_StableLM jeiku/Rosa_v1_3B jeiku/RPGPT_StableLM jeiku/Rosa_v1_3B jeiku/Toxic_DPO_StableLM jeiku/Rosa_v1_3B jeiku/Gnosis_256_StableLM jeiku/Rosa_v1_3B jeiku/Bluemoon_cleaned_StableLM
Merged Model	Yes
Model Size	3b
Required VRAM	5.6 GB
Updated	2026-07-30
Maintainer	jeiku
Model Type	stablelm_epoch
Model Files	5.6 GB: 1-of-1
Model Architecture	StableLMEpochForCausalLM
Context Length	4096
Model Max Length	4096
Transformers Version	4.35.2
Tokenizer Class	GPTNeoXTokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	50304
Torch Data Type	float16

Rank the Tofu 3B 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 Tofu 3B

Best Alternatives	Context / RAM	Downloads	Likes
Stable Code 3B Mlx	16K / 5.6 GB	17	1
Aura 3B	4K / 5.6 GB	3	2
Slim Extract	4K / 5.6 GB	10	13
Slim Boolean	4K / 5.6 GB	9	4
Slim Sa Ner	4K / 5.6 GB	20	6
Slim Tags 3B	4K / 5.6 GB	11	4
Slim Summary	4K / 5.6 GB	10	8
Slim Xsum	4K / 5.6 GB	14	6
Memphis CoT 3B	4K / 5.6 GB	24	31
Fett Uccine Mini 3B	4K / 5.6 GB	6	3

Note: green Score (e.g. "73.2") means that the model is better than jeiku/Tofu_3B.