CacheRouter: A Dual-Path Tool Routing Architecture with Cache-Preserving Main-Model Isolation for Long-Tail Tool Discovery
2026-08-24 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors explain a way to balance using many tools with keeping the prompt to a language model (LLM) small and reusable. They split tool handling into two parts: a fixed set of core tools seen by the main model, and a separate channel where another model picks and runs different tools as needed. This lets the main prompt stay stable so it can be cached and reused, reducing costs. Their method was tested and showed much lower input costs thanks to high cache hit rates.
Large Language ModelsTool UsePrompt EngineeringCachingProgressive DisclosureRoutingSub-modelToken-level CacheRequest ArchitectureCost Efficiency
Authors
Donghui Zha, Lingwei Xu, Linxiao Wu, Yixue Dong, Haochen Li
Abstract
Tool use in LLM systems faces a structural trade-off. Progressive disclosure keeps the prompt small by showing only the tools relevant to the current task, while prompt caching rewards a request prefix that stays fixed across calls; every change to the visible tool list invalidates the cached prefix. This paper treats the trade-off as a problem of request architecture and proposes a dual-path routing design that assigns tool selection and tool delivery to separate channels. The main model always sees a small, fixed set of core tools, so the head of its request is unchanged across calls; all other tools are reached through an independent routing channel, in which a router sub-model searches the full tool list, selects one tool, executes it, and returns the result. Tool registration is automated from source code and supports runtime updates, so the tool set can grow without modifying the main model's request prefix. The design generalizes progressive disclosure: capabilities are disclosed through the routing channel, and the main model's prefix stays stable. A prototype implementation was exercised on 55 functional queries and a 30-turn dialogue; token-level cache hit rates reached 90.99% and 95.2%, cutting input cost to about 12.0% and 8.0% of a no-cache baseline under DeepSeek's pricing, where cache-hit input tokens cost roughly 1/30 of cache-miss tokens.