iapp/chinda-qwen3-4b

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:4BQuant:BF16Ctx Length:32kPublished:May 28, 2025License:apache-2.0Architecture:Transformer0.0K Open Weights Warm

Chinda Opensource Thai LLM 4B by iApp Technology is a 4 billion parameter language model built on the Qwen3-4B architecture, specifically fine-tuned for the Thai language. It features advanced reasoning capabilities with a unique 'thinking mode' and achieves 98.4% accuracy in Thai language output, preventing foreign language interference. Optimized for resource efficiency and edge computing, it excels in RAG applications, mathematical calculations, and code assistance, making it ideal for sovereign AI solutions in Thailand.

Loading preview...

Chinda Opensource Thai LLM 4B: Sovereign AI for Thailand

Chinda Opensource Thai LLM 4B, developed by iApp Technology, is a 4 billion parameter model based on the Qwen3-4B architecture, designed to be a leading Thai language model. It is completely free and open-source under the Apache 2.0 License, allowing for commercial use, modification, and distribution.

Key Capabilities

  • Advanced Thinking Mode: Achieves the highest score among Thai LLMs in the 4B category, offering superior reasoning for complex problems and dynamic switching between thinking and non-thinking modes for efficiency.
  • Exceptional Thai Language Accuracy: Boasts 98.4% accuracy in Thai output, specifically fine-tuned to prevent unwanted foreign language generation.
  • Optimized Performance: Demonstrates 37% better overall performance compared to its nearest alternative in benchmarks like AIME24, LiveCodeBench, and MATH500, particularly strong in math and code tasks.
  • Resource Efficient: Designed for fast inference with minimal GPU memory consumption, suitable for mobile, laptop, and edge computing applications.
  • Extended Context: Supports a native context length of 32,768 tokens, extendable up to 131,072 tokens with YaRN scaling.

Good For

  • RAG Applications: Ideal for building Retrieval-Augmented Generation systems requiring data processing within Thai sovereignty.
  • Mathematical Calculation & Code Assistance: Excels in reasoning for math problems and providing programming support.
  • Mobile and Laptop Applications: Optimized as a Small Language Model for efficient deployment on personal devices.
  • Commercial Use: Its Apache 2.0 license permits broad commercial and research applications without restrictions.

Finetunes
3 models