What are the hardware requirements for Karasu Moe2?

Karasu Moe2 requires approximately 7.5 GB of VRAM and supports a context window of 2K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Karasu Moe2 and how large is it?

Karasu Moe2 is developed by aipib, a model with 1.1b 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 Karasu Moe2?

Karasu Moe2 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Karasu Moe2 by aipib — VRAM 7.5GB, 2K context

Name: Karasu Moe2
Author: aipib

Karasu Moe2 is an open-source language model by aipib. Features: 1.1b LLM, VRAM: 7.5GB, Context: 2K, License: apache-2.0, LLM Explorer Score: 0.12.

Base model:lightblue/karasu-1.... Base model:merge:lightblue/kar... Base model:merge:yuiseki/karas... Base model:yuiseki/karasu-sake... Endpoints compatible Frankenmoe Lazymergekit Lightblue/karasu-1.1b Merge Mergekit Mixtral Moe Region:us Safetensors Sharded Tensorflow Yuiseki/karasu-sake-qa-v0.1

Model Card on HF 🤗: https://huggingface.co/aipib/karasu-moe2

Karasu Moe2 Parameters and Internals

Model Type

moe, frankenmoe

Additional Notes

Karasu-moe2 is a Mixture of Experts model.

LLM Name	Karasu Moe2
Repository 🤗	https://huggingface.co/aipib/karasu-moe2
Base Model(s)	yuiseki/karasu-sake-qa-v0.1 lightblue/karasu-1.1B yuiseki/karasu-sake-qa-v0.1 lightblue/karasu-1.1B
Model Size	1.1b
Required VRAM	7.5 GB
Updated	2026-06-25
Maintainer	aipib
Model Type	mixtral
Model Files	4.0 GB: 1-of-2 3.5 GB: 2-of-2
Model Architecture	MixtralForCausalLM
License	apache-2.0
Context Length	2048
Model Max Length	2048
Transformers Version	4.40.2
Tokenizer Class	LlamaTokenizer
Padding Token	<s>
Vocabulary Size	32000
Torch Data Type	float32

Rank the Karasu Moe2 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 Karasu Moe2

Best Alternatives	Context / RAM	Downloads	Likes
TinyDolphin 3x MoE	4K / 5.3 GB	11	1
...inyLlama 1.1B Chat V1.0 X2 MoE	2K / 3.7 GB	2595	1
Tinyllama Moe4	2K / 3.7 GB	9	1
Tinyllama Moe1	2K / 5.3 GB	11	0
...yMoE DopeykarasuMoe Xdareties2	2K / 3.7 GB	6	0
Tinyllama Moe3	2K / 3.7 GB	5	0
LiteMOE 3x460m	1K / 2.1 GB	11	1

Note: green Score (e.g. "73.2") means that the model is better than aipib/karasu-moe2.