AgentWeave: Routing Before Reasoning for Efficient Function Calling in Tool-Rich Language Models
2026-08-24 • Artificial Intelligence
Artificial IntelligenceComputation and Language
AI summaryⓘ
The authors explore a way to help large language models pick the right tool to use by first narrowing down the choices before the model even starts figuring out what to do. They created AgentWeave, which acts like a filter that only shows the model a smaller set of relevant tools based on certain rules. When tested on specific tasks, AgentWeave helped the model succeed more often than when using all available tools or other methods. Although the overall success rates were still low, the results suggest that choosing smaller candidate tool sets can improve model performance without changing the model itself.
large language modelsfunction callingtool selectioncandidate action spaceroutingAgentWeaveBFCLtoken consumptionmodel latencybootstrap confidence interval
Authors
Saurav Singla, Aarav Singla, Advik Gupta, Parnika Gupta
Abstract
Large language models increasingly operate over large collections of tools, functions, APIs, and specialized agents. As the candidate action space grows, a function-calling model must process more schemas, consume more prompt tokens, and distinguish among increasingly similar or irrelevant alternatives. We study a complementary systems strategy: reduce the candidate set before language-model inference while leaving the downstream model unchanged. We introduce AgentWeave, a deterministic pre-inference routing layer that constructs a bounded model-visible action space using eligibility, requirement, capability, and routing signals. We evaluate AgentWeave with a frozen BFCL-derived routing-pressure protocol using the public MadeAgents/Hammer2.1-1.5b model. On 48 fresh BFCL V4 multiple-function tasks, AgentWeave achieves 6/48 (12.5%) native BFCL successes, whereas all-tools, deterministic random top-8, and semantic top-8 baselines each achieve 0/48. The paired success difference is +12.5 percentage points with a 10,000-resample paired bootstrap 95% confidence interval of +4.17 to +22.92 points and exact McNemar p=0.03125. Relative to all-tools exposure, AgentWeave presents 70.18% fewer tools, uses 61.70% fewer input tokens, and exhibits 50.95% lower mean local-model latency. The result is deliberately narrow: this is a BFCL-derived routing-pressure study rather than an official full BFCL leaderboard score, and absolute task success remains low. The evidence nevertheless shows that candidate-space construction can materially affect a fixed model's function-calling behavior and motivates evaluating routing as a distinct stage before model reasoning.