FutureWeaver: Planning Test-Time Compute for Multi-Agent Systems with Modularized Collaboration

TL;DR

FutureWeaver is a framework that optimizes test-time compute allocation in multi-agent systems, enabling better collaboration and performance under fixed compute budgets through modular workflows and dual-level planning.

Contribution

It introduces a novel modularized collaboration approach and a dual-level planning architecture for optimizing compute in multi-agent systems during inference.

Findings

Outperforms baselines across diverse budget settings

Effectively fosters collaboration among agents

Demonstrates scalability on complex benchmarks

Abstract

Scaling test-time computation improves large language model performance without additional training. Recent work demonstrates that techniques such as repeated sampling, self-verification, and self-reflection can significantly enhance task success by allocating more inference-time compute. However, applying these techniques across multiple agents in a multi-agent system is difficult: there does not exist principled mechanisms to allocate compute to foster collaboration among agents, to extend test-time scaling to collaborative interactions, or to distribute compute across agents under explicit budget constraints. To address this gap, we propose FutureWeaver, a framework for planning and optimizing test-time compute allocation in multi-agent systems under fixed budgets. FutureWeaver introduces modularized collaboration, formalized as callable functions that encapsulate reusable multi-agent workflows. These modules are automatically derived through self-play reflection by abstracting recurring interaction patterns from past trajectories. Building on these modules, FutureWeaver employs a dual-level planning architecture that optimizes compute allocation by reasoning over the current task state while also speculating on future steps. Experiments on complex agent benchmarks demonstrate that FutureWeaver consistently outperforms baselines across diverse budget settings, validating its effectiveness for multi-agent collaboration in inference-time optimization.