What are the hardware requirements for CAT Translate 0.8B?

CAT Translate 0.8B requires approximately 1.6 GB of VRAM and supports a context window of 8K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed CAT Translate 0.8B and how large is it?

CAT Translate 0.8B is developed by cyberagent, a model with 0.5b 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 CAT Translate 0.8B?

CAT Translate 0.8B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

CAT Translate 0.8B by cyberagent — VRAM 1.6GB, 8K context

Name: CAT Translate 0.8B
Author: cyberagent

CAT Translate 0.8B is an open-source language model by cyberagent. Features: 0.5b LLM, VRAM: 1.6GB, Context: 8K, License: mit, LLM Explorer Score: 0.23.

Arxiv:2509.22047 Arxiv:2606.21413 Base model:finetune:sbintuitio... Base model:sbintuitions/sarash... Dataset:cyberagent/cat-transla... En Ja Llama Region:us Safetensors

Model Card on HF 🤗: https://huggingface.co/cyberagent/CAT-Translate-0.8b

CAT Translate 0.8B Parameters and Internals

LLM Name	CAT Translate 0.8B
Repository 🤗	https://huggingface.co/cyberagent/CAT-Translate-0.8b
Base Model(s)	sbintuitions/sarashina2.2-0.5b sbintuitions/sarashina2.2-0.5b
Model Size	0.5b
Required VRAM	1.6 GB
Updated	2026-08-22
Maintainer	cyberagent
Model Type	llama
Model Files	1.6 GB
Supported Languages	ja en
Model Architecture	LlamaForCausalLM
License	mit
Context Length	8192
Model Max Length	8192
Transformers Version	4.57.6
Tokenizer Class	LlamaTokenizer
Padding Token	<pad>
Vocabulary Size	102400

Rank the CAT Translate 0.8B 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 CAT Translate 0.8B

Best Alternatives	Context / RAM	Downloads	Likes
Turn Detector	8K / 0.5 GB	694536	123
...arashina2.2 0.5B Instruct V0.1	8K / 1.6 GB	14012	15
...CLOVAX SEED Text Instruct 0.5B	8K / 1.1 GB	4195	85
Sarashina2.2 0.5B	8K / 1.6 GB	2049	16
Qwen1.5 0.5B Chat Llamafy	4K / 1.2 GB	239	5
Qwen2 0.5B Instruct Llamafy	4K / 1 GB	10	0
Npc Nano 0.5B Base	2K / 1 GB	8	0

Note: green Score (e.g. "73.2") means that the model is better than cyberagent/CAT-Translate-0.8b.