OpenThaiGPT 1.6 72b: Advanced Thai Language Model

OpenThaiGPT 1.6 72b is a 72-billion-parameter language model developed by the OpenThaiGPT Team, focusing on general-purpose tasks with strong reasoning capabilities in both Thai and English. This iteration significantly improves upon previous OpenThaiGPT versions, particularly in code reasoning and mathematical tasks, while maintaining high general language understanding.

Key Capabilities & Improvements

• Advanced Thai Language Processing: A 72-billion-parameter model with enhanced understanding of Thai language and cultural context.
• Strong Reasoning: Demonstrates robust reasoning capabilities in both Thai and English, crucial for complex problem-solving.
• Improved Code Reasoning: Shows significantly better performance in code reasoning benchmarks (LiveCodeBench-TH: 32.43, LiveCodeBench: 54.21) compared to OpenThaiGPT 1.5 72b.
• Balanced Performance: Achieves balanced capabilities across mathematical (MATH500: 82), coding, and general language tasks (OpenThaiEval: 78.7, Language Accuracy: 98.2).
• Technical Report: A detailed technical report is available at arXiv:2504.01789.

When to Use This Model

• Thai-centric Applications: Ideal for applications requiring deep understanding and generation of Thai language.
• Complex Reasoning Tasks: Suitable for use cases demanding strong reasoning in both Thai and English, including mathematical and coding challenges.
• General-Purpose LLM: A versatile model for a wide range of general language tasks where improved performance over previous versions is desired.