交接代價：LLM智能體中持續存在的非原生軌跡

摘要

編碼代理執行長期運行的任務，涵蓋數十次模型呼叫、工具使用及程式碼編輯。隨著這些運行逐步展開，使用者面臨一項實際的成本與品質之間的權衡：當較便宜的模型難以應付時升級至更強的模型，或在艱難的推理階段完成後降級。每次切換都要求接收方延續由另一模型所產生的非原生軌跡。我們研究此交接如何影響品質與成本，以及改變接收方所繼承的軌跡資訊如何左右最終結果。我們使用來自Claude和GPT系列的成對低成本、低能力（LC）與高成本、高能力（HC）模型，改變交接方向、時機及介面，並比較完整軌跡轉移、壓縮，以及在保留儲存庫狀態的前提下移除軌跡這幾種做法。在兩個模型系列中，完整軌跡升級僅能恢復LC到HC品質差距中不到一半的品質，卻同時帶來可觀的成本溢價。我們將此成本-品質懲罰稱之為「交接稅」。相較之下，降級則提供了有利的成本-品質平衡點。有趣的是，偏好的介面也隨交接方向而反轉：減少LC模型的軌跡資訊可提升升級品質，而移除HC模型的軌跡則會降低降級品質。

English

Coding agents perform long-running tasks spanning dozens of model calls, tool uses, and code edits. As these runs unfold, users face a practical cost-quality trade-off: escalating to a stronger model when a cheaper one struggles, or downshifting once the hard reasoning is complete. Each switch requires the receiver to continue a non-native trajectory produced by another model. We study how this handoff affects quality and cost, and how varying the trajectory information inherited by the receiver changes the outcome. Using pairs of low-cost, low-capability (LC) and high-cost, high-capability (HC) models from the Claude and GPT families, we vary handoff direction, timing, and interface, comparing full-trajectory transfer, compaction, and trajectory removal while preserving the repository state. Across both model families, full-trajectory escalation recovers less than half of the LC-to-HC quality gap while incurring a substantial cost premium. We term this cost-quality penalty the handoff tax. By contrast, downshift offers a favorable cost-quality point. Interestingly, the preferred interface also reverses with direction: reducing LC-model trajectory information improves escalation quality, whereas removing the HC-model trajectory reduces downshift quality.