Kanana 1.5 Open - source Bilingual Large Language Model: Strong in Programming and Mathematics, Supports Ultra

Kanana 1.5 2.1b Instruct 2505

Developed by kakaocorp

Kanana 1.5 is a bilingual large language model developed by Kakao, with significant improvements in programming, mathematics, and function calling capabilities. It supports a 32K token context length and can process ultra-long texts up to 128K tokens using YaRN extension technology.

Large Language Model

Transformers

Supports Multiple LanguagesOpen Source License:Apache-2.0 #Bilingual Instruction Optimization #Ultra-Long Context Processing #Programming & Math Enhancement

Downloads 1,361

Release Time : 5/21/2025

Model Overview

Kanana 1.5 is an efficient bilingual large language model focused on enhancing programming, mathematics, and function calling capabilities, suitable for complex scenarios and long-text processing.

Model Features

Enhanced Programming Capabilities

Excellent performance in programming benchmarks such as HumanEval and MBPP.

Mathematical Reasoning

High scores in mathematical benchmarks like GSM8K and MATH.

Long-Text Processing

Native support for 32K token context length, extendable to 128K tokens via YaRN.

Bilingual Support

Supports both English and Korean processing.

Optimized Dialogue Capabilities

Post-training optimization results in more natural and precise conversational responses.

Model Capabilities

Text Generation

Code Generation

Mathematical Problem Solving

Long-Text Comprehension

Bilingual Processing

Instruction Following

Use Cases

Programming Assistance

Code Generation

Generate code based on natural language descriptions.

Achieved 68.90 points in the HumanEval test.

Code Completion

Auto-complete code based on context.

Achieved 65.08 points in the MBPP+ test.

Mathematical Applications

Mathematical Problem Solving

Solve complex mathematical problems.

Achieved 81.43 points in the GSM8K zero-shot test.

Dialogue Systems

Intelligent Customer Service

Provide natural and fluent bilingual customer service dialogues.

Achieved 7.01 points in the MT-Bench test.

🚀 Kanana 1.5 - 2.1B Instruct 2505

Kanana 1.5 is a newly enhanced version of the Kanana model family, offering improved coding, math, and function - calling capabilities, and better handling of long - text scenarios.

🌟 1.5 HF Models | 📝 1.5 Blog | 📖 Technical Report

📢 News

🗞️ 2025/05/23: Published a blog post about Kanana 1.5 models and released 🌟HF model weights.
📖2025/02/27: Released Technical Report and 🌟HF model weights.
📝2025/01/10: Published a blog post about the development of Kanana Nano model.
📝2024/11/14: Published blog posts (pre - training, post - training) about the development of Kanana models.
🎥2024/11/06: Published a presentation video about the development of the Kanana models.

📚 Table of Contents

Kanana 1.5
- Performance
  - Base Model Evaluation
  - Instruct Model Evaluation
Contributors
Citation
Contact

🚀 Kanana 1.5

Kanana 1.5, a newly introduced version of the Kanana model family, presents substantial enhancements in coding, mathematics, and function calling capabilities over the previous version, enabling broader application to more complex real - world problems. This new version now can handle up to 32K tokens length natively and up to 128K tokens using YaRN, allowing the model to maintain coherence when handling extensive documents or engaging in extended conversations. Furthermore, Kanana 1.5 delivers more natural and accurate conversations through a refined post - training process.

⚠️ Important Note

Neither the pre - training nor the post - training data includes Kakao user data.

✨ Performance

🔍 Base Model Evaluation

Models	MMLU	KMMLU	HAERAE	HumanEval	MBPP	GSM8K
Kanana - 1.5 - 2.1B	56.30	45.10	77.46	52.44	47.00	55.95
Kanana - Nano - 2.1B	54.83	44.80	77.09	31.10	46.20	46.32

🔍 Instruct Model Evaluation

Models	MT - Bench	KoMT - Bench	IFEval	HumanEval+	MBPP+	GSM8K (0 - shot)	MATH	MMLU (0 - shot, CoT)	KMMLU (0 - shot, CoT)	FunctionChatBench
Kanana - 1.5 - 2.1B*	7.01	6.54	68.61	68.90	65.08	81.43	60.62	53.87	32.93	53.70
Kanana - Nano - 2.1B	6.40	5.90	71.97	63.41	62.43	72.32	29.26	52.48	38.51	26.10

⚠️ Important Note

* Models released under Apache 2.0 are trained on the latest versions compared to other models.

👥 Contributors

Language Model Training: Yunju Bak, Doohae Jung, Boseop Kim, Nayeon Kim, Hojin Lee, Jaesun Park, Minho Ryu
Language Model Alignment: Jiyeon Ham, Seungjae Jung, Hyunho Kim, Hyunwoong Ko, Changmin Lee, Daniel Wontae Nam
AI Engineering: Youmin Kim, Hyeongju Kim

📄 Citation

@misc{kananallmteam2025kananacomputeefficientbilinguallanguage,
      title={Kanana: Compute-efficient Bilingual Language Models}, 
      author={Kanana LLM Team and Yunju Bak and Hojin Lee and Minho Ryu and Jiyeon Ham and Seungjae Jung and Daniel Wontae Nam and Taegyeong Eo and Donghun Lee and Doohae Jung and Boseop Kim and Nayeon Kim and Jaesun Park and Hyunho Kim and Hyunwoong Ko and Changmin Lee and Kyoung-Woon On and Seulye Baeg and Junrae Cho and Sunghee Jung and Jieun Kang and EungGyun Kim and Eunhwa Kim and Byeongil Ko and Daniel Lee and Minchul Lee and Miok Lee and Shinbok Lee and Gaeun Seo},
      year={2025},
      eprint={2502.18934},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2502.18934}, 
}

📞 Contact

Kanana LLM Team Technical Support: kanana - llm@kakaocorp.com
Business & Partnership Contact: alpha.k@kakaocorp.com

📄 License

This project is licensed under the Apache 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご