What are the hardware requirements for ATOMIS?

ATOMIS requires approximately 66 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed ATOMIS and how large is it?

ATOMIS is developed by KINS-ai, a model with 27b 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 ATOMIS?

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

ATOMIS by KINS-ai — VRAM 66GB, 32K context

Name: ATOMIS
Author: KINS-ai

ATOMIS is an open-source language model by KINS-ai. Features: 27b LLM, VRAM: 66GB, Context: 32K, License: gemma, LLM Explorer Score: 0.17.

Arxiv:2310.10920 Arxiv:2408.01262 Base model:finetune:google/gem... Base model:google/gemma-2-27b Conversational En Endpoints compatible Gemma-2 Gemma2 Kins-ai Ko Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/KINS-ai/ATOMIS

ATOMIS Parameters and Internals

LLM Name	ATOMIS
Repository 🤗	https://huggingface.co/KINS-ai/ATOMIS
Base Model(s)	Gemma 2 27B google/gemma-2-27b
Model Size	27b
Required VRAM	66 GB
Updated	2026-07-05
Maintainer	KINS-ai
Model Type	gemma2
Model Files	4.7 GB: 1-of-14 4.9 GB: 2-of-14 4.9 GB: 3-of-14 5.0 GB: 4-of-14 4.9 GB: 5-of-14 4.9 GB: 6-of-14 5.0 GB: 7-of-14 4.9 GB: 8-of-14 4.9 GB: 9-of-14 5.0 GB: 10-of-14 4.9 GB: 11-of-14 4.9 GB: 12-of-14 5.0 GB: 13-of-14 2.1 GB: 14-of-14
Supported Languages	en ko
Model Architecture	Gemma2ForCausalLM
License	gemma
Context Length	32768
Model Max Length	32768
Transformers Version	4.43.4
Tokenizer Class	GemmaTokenizer
Padding Token	<pad>
Vocabulary Size	256000
Torch Data Type	bfloat16

Rank the ATOMIS 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 ATOMIS

Best Alternatives	Context / RAM	Downloads	Likes
SystemGemma2 27B It	32K / 54.7 GB	96	1
SystemGemma2 27B It	32K / 54.7 GB	11	3
Gemma 2 27B It	8K / 54.7 GB	117753	568
Gemma 2 27B It	8K / 54.7 GB	52166	6
...izon AI Korean Avengers V3 27B	8K / 54.8 GB	8	0
...izon AI Korean Avengers V2 27B	8K / 54.8 GB	8	0
Gemma2Slerp2 27B	8K / 54.8 GB	14	1
...s Horizon AI Korean Superb 27B	8K / 54.8 GB	8	2
GemmaStock1 27B	8K / 54.8 GB	8	0
Gemma2Slerp3 27B	8K / 54.8 GB	17	0

Note: green Score (e.g. "73.2") means that the model is better than KINS-ai/ATOMIS.