CITATION.cff

cff-version: 1.2.0
title: "Erk-14B: A Turkish language model"
message: "If you use this model, please cite it as below."
type: software
authors:
  - name: "eCloud Tech."
    website: "https://www.e-cloud.web.tr"
    city: "Şanlıurfa"
    country: TR
repository-code: "https://github.com/ecloudtechnology/erk"
url: "https://huggingface.co/ecloudtech/Erk-14B"
abstract: >-
  Erk-14B is a Turkish language model built on Qwen3-14B through Turkish
  continued pretraining and instruction tuning on real documents. It scores
  69.7% on TurkishMMLU (0-shot, 9 subjects), raising the base model from 63.4%
  and ranking first among open Turkish models at the time of release. The model
  targets Turkish morphology — root and suffix structure — rather than treating
  Turkish as a foreign language to be segmented character by character.
keywords:
  - turkish-nlp
  - large-language-models
  - turkish
  - qwen3
  - instruction-tuning
  - continued-pretraining
license-url: "https://github.com/ecloudtechnology/erk/blob/main/LICENSE"
version: "1.0"
date-released: "2026-07-29"