Agentic Routing: The Harness-Native Data Flywheel

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Xinchen Liu, Hang Zhou, Yingjie Zong, Yuchuan Tian, Liuyang Song, Shuo Zhang, Yulong Li, Wei He, Mengyu Zheng, Runke Liu, Siyang Cheng, Xiang Kuang, Hailin Hu, Kai Han, Yunhe Wang
• Venue: arXiv: cs.CL
• Published: 2026-07-13
• Collected: 2026-07-14
• URL: http://arxiv.org/abs/2607.11399v1
• PDF: https://arxiv.org/pdf/2607.11399v1
• DOI: UNKNOWN

Plain-Language Brief

The authors propose a new method of model selection for large language model agents called Harness-Native agentic routing. This method selects the best-fit model or multiple complementary models based on the full harness state, creating a data flywheel that improves cost-quality trade-offs and generates more training data.

What the Document Says

• Large language model agents are managed by an execution harness.
• Existing routing methods optimize single-turn cost-quality trade-offs but miss the unique aspects of agents (execution state, intermediate failures, and feedback loops).
• Proposed is Harness-Native agentic routing, a step-level routing paradigm that selects a single best-fit model or multiple complementary models based on the full harness state.
• Every routing decision produces a structured data record consisting of various elements.
• These records form a harness-native data flywheel: execution traces train better routers and harness-native models, which improve cost-quality trade-offs and generate more traces under the same budget.
• The idea is instantiated in OpenSquilla with a four-layer routing stack, an open LightGBM cold-start ranker, and a staged router-model path that turns logged arena records into progressively stronger routing policies.
• The report studies singleton and multi-model routing on agentic benchmarks including DRACO and PinchBench.

Key Concepts Mentioned

• large language model agents
• execution harness
• routing methods
• cost-quality trade-offs
• Harness-Native agentic routing
• data flywheel
• OpenSquilla
• DRACO
• PinchBench

Claims Made by the Document

• Existing routing methods are inadequate for agents.
• Harness-Native agentic routing improves cost-quality trade-offs and generates more training data.
• OpenSquilla is an instantiation of the Harness-Native idea.

Evidence Used

• The text describes the limitations of existing routing methods and presents Harness-Native agentic routing as a solution.
• Details about OpenSquilla and its components are provided.

Methods / Approach

UNKNOWN

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

Large language model agents are increasingly executed not by a single model call, but by an execution harness that manages observation, context, control, action, state, and verification. At the same time, frontier and open models are becoming structurally specialized: a model that is strong at code editing, long-context recovery, tool use, mathematical reasoning, or low-latency response may not dominate on the other axes. This makes model selection inside an agent a core systems problem rather than a per-query serving trick. Existing routing methods mostly optimize single-turn cost-quality trade-offs and therefore miss the execution state, intermediate failures, and feedback loops that make agents different from chat completion. We propose Harness-Native agentic routing, a step-level routing paradigm that selects either a single best-fit model for cost-effective execution or multiple complementary models for ensemble-style accuracy improvement, conditioned on the full harness state. The key insight is that every routing decision naturally produces a structured data record -- consisting of the query, harness state, model choice or model set, execution trace, outcome, and cost -- whose labels are supplied by the environment rather than by the router itself. These records form a harness-native data flywheel: execution traces train better routers and harness-native models, which improve cost-quality trade-offs and generate more traces under the same budget. We instantiate this idea in OpenSquilla with a four-layer routing stack, an open LightGBM cold-start ranker, and a staged router-model path that turns logged arena records into progressively stronger routing policies. The report studies singleton and multi-model routing on agentic benchmarks including DRACO and PinchBench, and argues that agentic routing is not merely cost control, but a data engine for agent-native training.