Metraipolis › archive › dc_c8b4fe41ea7e

AgentRouter: Heterogeneous Model Routing for Cost-Optimal Multi-Step Agentic Workflows — arXiv:2609.22951

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.22951
License (as archived): CC0-1.0
Archived: 2026-10-09T06:50:30.833Z
SHA-256 of the archived text: 7370400018d71be80424e6b6c51706d8888d4635fbca57a7665d15bf4d0b3027
Stable URL of this page: https://metraipolis.com/documento/dc_c8b4fe41ea7e

How to cite

AgentRouter: Heterogeneous Model Routing for Cost-Optimal Multi-Step Agentic Workflows — arXiv:2609.22951 — arXiv (en), https://arxiv.org/abs/2609.22951 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c8b4fe41ea7e (Metraipolis archive dc_c8b4fe41ea7e, sha256 7370400018d71be80424e6b6c51706d8888d4635fbca57a7665d15bf4d0b3027)

BibTeX

@misc{dcc8b4fe41ea7e,
  title = {AgentRouter: Heterogeneous Model Routing for Cost-Optimal Multi-Step Agentic Workflows — arXiv:2609.22951},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.22951},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c8b4fe41ea7e; sha256 7370400018d71be80424e6b6c51706d8888d4635fbca57a7665d15bf4d0b3027},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.22951. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Enterprise agentic systems that route every trajectory step to a frontier model waste 60-80% of their inference budget on subtasks that smaller models handle equally well. Existing routing solutions optimize single-turn query assignment but ignore a property unique to agentic workflows: subtask complexity varies widely within a single trajectory. A planning step may require frontier-class reasoning while a subsequent formatting step needs only a 7B model. We formalize step-level model routing as a sequential assignment problem over agent trajectories and propose AgentRouter, a lightweight classifier (12M parameters, <5ms overhead per step on an A100 GPU) that maps each trajectory step to one of four model tiers using five features extractable at routing time. Trained on 50,000 annotated agent trajectory steps spanning planning, coding, research, and data analysis tasks, AgentRouter achieves 72% cost reduction relative to frontier-only baselines, retaining 97.3% of frontier-only quality (less than 3% degradation in end-to-end task completion); per-step routing accuracy reaches 91% on minimal-complexity steps and 85% on efficient-tier steps, with 76-82% on the harder mid-range and frontier tiers. On the same benchmarks, RouteLLM and FrugalGPT (applied per-step) achieve only 31% and 44% cost reduction respectively, because their single-turn training signal misses trajectory-level quality dependencies.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-22 in cs.CL).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Machine Learning (cs.LG); Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2609.22951 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-09T06:48:45.111Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).