Pathumma
Open-source Thai LLM family from NECTEC for government and enterprise use
Pathumma is an open-source Thai LLM family best for Thai government agencies, universities, and local fintechs that need a self-hosted Thai LLM with strong coverage of legal, government, and academic text. Its SEA edge is being developed by NECTEC under NSTDA with permissive open licensing, making it the natural choice for Thai public-sector deployments where vendor lock-in is unacceptable and on-prem inference is required. The honest caveat: open weights with no managed SaaS means you carry the inference, fine-tuning, and ops burden - this is not a plug-and-play API.
- ✓Trained on Thai government, legal, and academic corpora for public-sector fit
- ✓Open weights on Hugging Face under permissive licensing avoid vendor lock-in
- ✓Self-hostable on a single A100 or RTX 4090 for inference
- ✓Compatible with vLLM, TGI, llama.cpp, and standard inference stacks
- ✓Backed by NECTEC and Thai government research funding for continuity
- ×Open weights mean you carry inference, ops, and fine-tuning burden
- ×No managed SaaS, billing, or support contracts available
- ×Smaller commercial ecosystem than Typhoon from SCB 10X
- ×Best for ML-capable teams, not no-code or business-user buyers
About Pathumma
Pathumma is an open-source Thai large language model family developed by NECTEC under Thailand's National Science and Technology Development Agency (NSTDA). Released free under permissive licensing on Hugging Face, Pathumma is trained on Thai legal, government, and academic text and is the default starting point for Thai government and university teams that need a self-hosted Thai LLM.
Key Features
Best For
Southeast Asia Fit
Pathumma is one of the few Thai LLMs released by a government research institution rather than a commercial entity, which makes it the natural choice for Thai public-sector deployments and universities. Compared to Typhoon (SCB 10X), Pathumma has stronger coverage of Thai legal and government text, and the open licensing avoids vendor lock-in concerns at ministries and state enterprises.
- Global
- SEA
- Thailand
Integrations
- Other
- Hugging Face
- Transformers
- PyTorch
- Safetensors
- GGUF
- text-generation-inference
This profile was researched and written with AI from Pathumma's public pages, and AI assistants quote pages on this site. Something wrong or out of date? Email info@software-listing.com with this page and a public source. Corrections are free, for every listing.
How listings and corrections work →Related Analysis & Guides
SEA AI Cost Optimization 2026: Self-Host vs API for Bahasa, Thai, Vietnamese Workloads
AI Agritech Tools for SEA Farmers in 2026: What's Actually Worth Using
AI Coding Assistants for SEA Developers in 2026: Cursor, Copilot, and What's Actually Worth Paying For
The questions operators actually ask.
Should Thai government agencies pick Pathumma or Typhoon?
It depends. Pathumma has stronger coverage of Thai legal and government text and avoids vendor lock-in concerns at ministries and state enterprises. Typhoon from SCB 10X has a richer commercial ecosystem. For public-sector use, Pathumma is usually the natural first pick.
Is Pathumma free to use commercially?
Yes. Pathumma is released under permissive open licensing on Hugging Face, which makes it free to self-host for commercial use. You pay only for your own inference infrastructure and engineering time.
Can a non-ML team deploy Pathumma in production?
No, not easily. Open weights without a managed SaaS mean you need ML engineering capability to handle inference, fine-tuning, and ops. Non-ML teams should pair it with an inference partner or use a hosted alternative instead.