REDSearcher: A Scalable and Cost-Efficient Framework for Long-Horizon Search Agents

TL;DR

REDSearcher introduces a comprehensive framework that enhances long-horizon search agents by improving task synthesis, tool use, and training efficiency, achieving state-of-the-art results in both text and multimodal benchmarks.

Contribution

It presents novel methods for scalable task generation, tool-augmented queries, and efficient training, addressing key challenges in long-horizon search agent development.

Findings

Achieves state-of-the-art performance on search benchmarks.

Provides a large dataset of high-quality search trajectories.

Reduces training costs through local simulation environment.

Abstract

Large language models are transitioning from generalpurpose knowledge engines to realworld problem solvers, yet optimizing them for deep search tasks remains challenging. The central bottleneck lies in the extreme sparsity of highquality search trajectories and reward signals, arising from the difficulty of scalable longhorizon task construction and the high cost of interactionheavy rollouts involving external tool calls. To address these challenges, we propose REDSearcher, a unified framework that codesigns complex task synthesis, midtraining, and posttraining for scalable searchagent optimization. Specifically, REDSearcher introduces the following improvements: (1) We frame task synthesis as a dualconstrained optimization, where task difficulty is precisely governed by graph topology and evidence dispersion, allowing scalable generation of complex, highquality tasks. (2) We introduce toolaugmented queries to encourage proactive tool use rather than passive recall.(3) During midtraining, we strengthen core atomic capabilities knowledge, planning, and function calling substantially reducing the cost of collecting highquality trajectories for downstream training. (4) We build a local simulated environment that enables rapid, lowcost algorithmic iteration for reinforcement learning experiments. Across both textonly and multimodal searchagent benchmarks, our approach achieves stateoftheart performance. To facilitate future research on longhorizon search agents, we will release 10K highquality complex text search trajectories, 5K multimodal trajectories and 1K text RL query set, and together with code and model checkpoints.