# When to Compile a Computer-Use Agent? Measuring Payback and Making Compilation Decisions for Token Efficiency

arXiv: https://arxiv.org/abs/2610.02932 · HTML: https://mlsanity.com/paper/2610.02932/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use.
- Builds on 20 of its 24 references that are in the map.
- First posted on arXiv 2 Oct 2026.

## Placed in

- LLM agents and tool use / [Software engineering coding agents](https://mlsanity.com/agents/software-engineering-coding-agents/)

## Builds on (most cited first, 20 of 20)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [WebArena: A Realistic Web Environment for Building Autonomous Agents](https://mlsanity.com/paper/2307.13854/) (2023)
- [OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments](https://mlsanity.com/paper/2404.07972/) (2024)
- [CogAgent: A Visual Language Model for GUI Agents](https://mlsanity.com/paper/2312.08914/) (2023)
- [AndroidWorld: A Dynamic Benchmarking Environment for Autonomous Agents](https://mlsanity.com/paper/2405.14573/) (2024)
- [Agent Workflow Memory](https://mlsanity.com/paper/2409.07429/) (2024)
- [Synapse: Trajectory-as-Exemplar Prompting with Memory for Computer Control](https://mlsanity.com/paper/2306.07863/) (2023)
- [UFO: A UI-Focused Agent for Windows OS Interaction](https://mlsanity.com/paper/2402.07939/) (2024)
- [Inducing Programmatic Skills for Agentic Tasks](https://mlsanity.com/paper/2504.06821/) (2025)
- [Improving Online Algorithms via ML Predictions](https://mlsanity.com/paper/2407.17712/) (2024)
- [Conservative Contextual Linear Bandits](https://mlsanity.com/paper/1611.06426/) (2016)
- [Benchmarking Mobile Device Control Agents across Diverse Configurations](https://mlsanity.com/paper/2404.16660/) (2024)
- [OpenApps: Simulating Environment Variations to Measure UI-Agent Reliability](https://mlsanity.com/paper/2511.20766/) (2025)
- [D-GARA: A Dynamic Benchmarking Framework for GUI Agent Robustness in Real-World Anomalies](https://mlsanity.com/paper/2511.16590/) (2025)
- [SkillSmith: Compiling Agent Skills into Boundary-Guided Runtime Interfaces](https://mlsanity.com/paper/2605.15215/) (2026)
- [PreAct: Computer-Using Agents that Get Faster on Repeated Tasks](https://mlsanity.com/paper/2606.17929/) (2026)
- [TraceCompiler: Skill-Guided Mining and Compilation of LLM Agent Traces into Mostly Deterministic Workflows](https://mlsanity.com/paper/2608.02680/) (2026)
- [Compiled AI: Deterministic Code Generation for LLM-Based Workflow Automation](https://mlsanity.com/paper/2604.05150/) (2026)
- [AutoRPA: Efficient GUI Automation through LLM-Driven Code Synthesis from Interactions](https://mlsanity.com/paper/2605.21082/) (2026)
- [Beyond Static Endpoints: Tool Programs as an Interface for Flexible Agentic Web Services](https://mlsanity.com/paper/2606.19992/) (2026)

## Abstract (arXiv)

Compiling GUI procedures that agents execute repeatedly into programs can reduce their token costs. However, measuring payback and deciding when to compile have two challenges. First, compilation costs are uncertain because attempts can require repair and still fail to produce a usable program. Second, future reuse is unknown because tasks may stop arriving or GUI drift may stop the program from working. To address these challenges, we propose PACE (Payback-Aware Compilation from Experience), a system with a measurement protocol and an online compilation algorithm. The measurement protocol records successful and failed compilation costs, and compares agent and program execution costs on matched task inputs to estimate per-use savings and payback counts. Using these measurements, the online algorithm compares estimated future savings with compilation costs, including failed attempts, based on past task arrivals and compilation outcomes. It checks execution and compilation charges against a cumulative budget determined by observed task arrivals before allowing either action. Under stated action-cost assumptions, total cost after each arrival is at most 1+ε times the cost of running every task with the agent. For successful compilation attempts, estimated payback counts excluding source agent runs are 2-16 uses. In simulations using recorded task arrivals, PACE reduces token costs by 17.3% compared with ReAct, 24.9% with the AutoRPA adaptation, and 17.3% with the ToolPro adaptation on average (ε=0.25).