[2601.04426] XGrammar-2: Efficient Dynamic Structured Generation Engine for Agentic LLMs

Computer Science > Artificial Intelligence arXiv:2601.04426 (cs) [Submitted on 7 Jan 2026 (v1), last revised 26 Mar 2026 (this version, v2)] Title:XGrammar-2: Efficient Dynamic Structured Generation Engine for Agentic LLMs Authors:Linzhang Li, Yixin Dong, Guanjie Wang, Ziyi Xu, Alexander Jiang, Tianqi Chen View a PDF of the paper titled XGrammar-2: Efficient Dynamic Structured Generation Engine for Agentic LLMs, by Linzhang Li and 5 other authors View PDF HTML (experimental) Abstract:Modern LLM agents increasingly rely on dynamic structured generation, such as tool calling and response protocols. Unlike traditional structured generation with static structures, these workloads vary both across requests and within a request, posing new challenges to existing engines. We present XGrammar-2, a structured generation engine for dynamic agentic workloads. Our design is based on two key ideas: first-class support for tag-triggered structure switching, and fine-grained reuse across requests with different output structures. Concretely, XGrammar-2 introduces TagDispatch for dynamic structural dispatching and Cross-Grammar Cache for substructure-level cache reuse across grammars. It further improves efficiency with an Earley-based adaptive token mask cache, just-in-time compilation, and repetition state compression. Experiments show that XGrammar-2 achieves over 6x faster compilation than prior structured generation engines, and incurs near-zero end-to-end overhead in modern LLM serv...