{"ID":2877722,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.20019","arxiv_id":"2508.20019","title":"Symphony: A Decentralized Multi-Agent Framework for Scalable Collective Intelligence","abstract":"Most existing Large Language Model (LLM)-based agent frameworks rely on centralized orchestration, incurring high deployment costs, rigid communication topologies, and limited adaptability. To address these challenges, we introduce Symphony, a decentralized multi-agent system which enables lightweight LLMs on consumer-grade GPUs to coordinate. Symphony introduces three key mechanisms: (1) a decentralized ledger that records capabilities, (2) a Beacon-selection protocol for dynamic task allocation, and (3) weighted result voting based on CoTs. This design forms a privacy-saving, scalable, and fault-tolerant orchestration with low overhead. Empirically, Symphony outperforms existing baselines on reasoning benchmarks, achieving substantial accuracy gains and demonstrating robustness across models of varying capacities.","short_abstract":"Most existing Large Language Model (LLM)-based agent frameworks rely on centralized orchestration, incurring high deployment costs, rigid communication topologies, and limited adaptability. To address these challenges, we introduce Symphony, a decentralized multi-agent system which enables lightweight LLMs on consumer-...","url_abs":"https://arxiv.org/abs/2508.20019","url_pdf":"https://arxiv.org/pdf/2508.20019v1","authors":"[\"Ji Wang\",\"Kashing Chen\",\"Xinyuan Song\",\"Ke Zhang\",\"Lynn Ai\",\"Eric Yang\",\"Bill Shi\"]","published":"2025-08-27T16:27:57Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\",\"cs.CL\",\"cs.MA\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}