What are the hardware requirements for K2S3 14B V0.2?

K2S3 14B V0.2 requires approximately 28.7 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 K2S3 14B V0.2 and how large is it?

K2S3 14B V0.2 is developed by Changgil, a model with 14b 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 K2S3 14B V0.2?

K2S3 14B V0.2 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

K2S3 14B V0.2 by Changgil — VRAM 28.7GB, 32K context

Name: K2S3 14B V0.2
Author: Changgil

K2S3 14B V0.2 is an open-source language model by Changgil. Features: 14b LLM, VRAM: 28.7GB, Context: 32K, License: cc-by-nc-4.0, LLM Explorer Score: 0.17.

En Endpoints compatible Ko Mistral Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Changgil/K2S3-14b-v0.2

K2S3 14B V0.2 Parameters and Internals

Additional Notes

The model has been enhanced with Korean vocabulary and merges to the tokenizer.

Supported Languages

en (English), ko (Korean)

Training Details

Data Sources:

alpaca-gpt4-data, The OpenOrca Dataset

Methodology:

Depth up scaling and full parameter tuning with SFT (Supervised Fine-Tuning)

Hardware Used:

A100 (80G*2EA) GPUs

Model Architecture:

Enhanced base model with depth up scaling

LLM Name	K2S3 14B V0.2
Repository 🤗	https://huggingface.co/Changgil/K2S3-14b-v0.2
Model Size	14b
Required VRAM	28.7 GB
Updated	2026-08-22
Maintainer	Changgil
Model Type	mistral
Model Files	5.0 GB: 1-of-6 4.9 GB: 2-of-6 5.0 GB: 3-of-6 4.9 GB: 4-of-6 4.9 GB: 5-of-6 4.0 GB: 6-of-6
Supported Languages	en ko
Model Architecture	MistralForCausalLM
License	cc-by-nc-4.0
Context Length	32768
Model Max Length	32768
Transformers Version	4.38.0.dev0
Tokenizer Class	LlamaTokenizer
Padding Token	<unk>
Vocabulary Size	48000
Torch Data Type	float16

Rank the K2S3 14B V0.2 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 K2S3 14B V0.2

Best Alternatives	Context / RAM	Downloads	Likes
...ral Nemo Instruct 14B Merge V1	1000K / 24.6 GB	19	0
ColdBrew 12B NemoMix Test0	128K / 24.5 GB	6	0
Wendigo 14B Alpha4	32K / 28.4 GB	1288	0
Qwen1.5 14B Chat Mistral	32K / 28.6 GB	11	4
Mistral 14B Merge Base	32K / 28.4 GB	92	2
Synthetic Minstrel 14B	32K / 27.6 GB	8	3
Wandering Minstrel 14B	32K / 27.6 GB	6	3
Phi 3 Medium 4K Instruct	4K / 28 GB	38	28
Barcenas 14B Phi 3 Medium ORPO	4K / 28 GB	8277	5
SauerkrautLM Phi 3 Medium	4K / 28 GB	8249	9

Note: green Score (e.g. "73.2") means that the model is better than Changgil/K2S3-14b-v0.2.