An Industrial-Scale Insurance LLM Achieving Verifiable Domain Mastery and Hallucination Control without Competence Trade-offs

TL;DR

This paper introduces INS-S1, an insurance-specific LLM trained with novel data synthesis and curriculum frameworks, achieving state-of-the-art domain performance and minimal hallucinations without losing general capabilities.

Contribution

The paper presents a new end-to-end alignment paradigm for insurance LLMs, combining hierarchical data synthesis and a progressive curriculum to ensure domain mastery and hallucination control.

Findings

INS-S1 achieves state-of-the-art performance on insurance tasks.

It maintains top-tier general capabilities.

It records a 0.6% hallucination rate, the lowest to date.

Abstract

Adapting Large Language Models (LLMs) to high-stakes vertical domains like insurance presents a significant challenge: scenarios demand strict adherence to complex regulations and business logic with zero tolerance for hallucinations. Existing approaches often suffer from a Competency Trade-off - sacrificing general intelligence for domain expertise - or rely heavily on RAG without intrinsic reasoning. To bridge this gap, we present INS-S1, an insurance-specific LLM family trained via a novel end-to-end alignment paradigm. Our approach features two methodological innovations: (1) A Verifiable Data Synthesis System that constructs hierarchical datasets for actuarial reasoning and compliance; and (2) A Progressive SFT-RL Curriculum Framework that integrates dynamic data annealing with a synergistic mix of Verified Reasoning (RLVR) and AI Feedback (RLAIF). By optimizing data ratios and reward signals, this framework enforces domain constraints while preventing catastrophic forgetting. Additionally, we release INSEva, the most comprehensive insurance benchmark to date (39k+ samples). Extensive experiments show that INS-S1 achieves SOTA performance on domain tasks, significantly outperforming DeepSeek-R1 and Gemini-2.5-Pro. Crucially, it maintains top-tier general capabilities and achieves a record-low 0.6% hallucination rate (HHEM). Our results demonstrate that rigorous domain specialization can be achieved without compromising general intelligence.