Typhoon-S: Minimal Open Post-Training for Sovereign Large Language Models

TL;DR

Typhoon-S presents a minimal, open post-training recipe for sovereign LLMs, addressing the need for regional control over weights and data under limited resources. By combining a lightweight two-stage adoptability pipeline (SFT followed by on-policy distillation) with a sovereignty-focused RFT extension (InK-GRPO) and agentic retrieval, the approach achieves strong Thai-language instruction-following while preserving broad capabilities. Empirical evaluations show SFT+OPD improves robustness and language coverage beyond SFT alone, and InK-GRPO enhances domain-specific Thai legal reasoning and agentic tasks without catastrophic forgetting. The work demonstrates that carefully designed post-training, using open data and moderate compute, can enable high-quality sovereign LLMs suitable for government and academic contexts, with explicit, transparent pipelines and releases. Overall, Typhoon-S provides practical recipes and open resources to enable sovereign adoption and jurisdiction-specific reasoning in resource-constrained settings.

Abstract

Large language models (LLMs) have progressed rapidly; however, most state-of-the-art models are trained and evaluated primarily in high-resource languages such as English and Chinese, and are often developed by a small number of organizations with access to large-scale compute and data. This gatekeeping creates a practical barrier for sovereign settings in which a regional- or national-scale institution or domain owner must retain control and understanding of model weights, training data, and deployment while operating under limited resources and strict transparency constraints. To this end, we identify two core requirements: (1) adoptability, the ability to transform a base model into a general-purpose assistant, and (2) sovereign capability, the ability to perform high-stakes, region-specific tasks (e.g., legal reasoning in local languages and cultural knowledge). We investigate whether these requirements can be achieved without scaling massive instruction corpora or relying on complex preference tuning pipelines and large-scale reinforcement fine-tuning (RFT). We present Typhoon S, a minimal and open post-training recipe that combines supervised fine-tuning, on-policy distillation, and small-scale RFT. Using Thai as a representative case study, we demonstrate that our approach transforms both sovereign-adapted and general-purpose base models into instruction-tuned models with strong general performance. We further show that small-scale RFT with InK-GRPO -- an extension of GRPO that augments the GRPO loss with a next-word prediction loss -- improves Thai legal reasoning and Thai-specific knowledge while preserving general capabilities. Our results suggest that a carefully designed post-training strategy can reduce the required scale of instruction data and computation, providing a practical path toward high-quality sovereign LLMs under academic-scale resources.

Figures (3)

• Figure 1: Overview of the target-language dataset construction pipeline for Thai.
• Figure 2: Prompt template for calculating RFT rewards using LLM-as-a-judge evaluation.
• Figure 3: LLM-judge prompt used to compute an accuracy score by comparing model responses against a ground-truth answer.