WorkstreamBench: Evaluating LLM Agents on End-to-End Spreadsheet Tasks in Finance

2026-05-21

AI Summary

Existing benchmarks inadequately assess the ability of large language model (LLM) agents to generate complete spreadsheets end-to-end in financial contexts—such as financial modeling or scenario analysis—being largely confined to question answering or single-formula editing. This work proposes the first end-to-end spreadsheet generation evaluation framework tailored to real-world financial workflows, evaluating performance along three core dimensions: accuracy, formula correctness, and formatting compliance, while also introducing professional criteria such as readability and modifiability for the first time. The framework integrates both human and automated evaluation methods for comprehensive assessment. Experimental results demonstrate that Claude-family models achieve the strongest performance among current LLMs, yet still fall significantly short of expert human practitioners on complex tasks, revealing critical limitations of contemporary LLM agents in authentic financial settings.

Abstract

LLM agents are increasingly expected to carry out end-to-end workflows, producing complete artifacts from high-level user instructions. To meet enterprise needs, frontier AI labs have developed agents that can construct entire spreadsheets from scratch. This is especially relevant in finance, where core workflows such as financial modeling, forecasting, and scenario analysis are commonly conducted through spreadsheets. Yet, existing spreadsheet benchmarks do not measure this advanced capability, focusing instead on question-answering or single-formula edits. To address this gap, we provide one of the first evaluations of agents on end-to-end spreadsheet tasks, focusing on economically critical financial workflows such as modeling and scenario analysis. Since deliverables therein are routinely reviewed and revised by multiple stakeholders, judging their quality necessarily involves high-level criteria such as readability or ease of modification. To reflect the multidimensional nature of solution quality, we develop an evaluation taxonomy comprising three dimensions: Accuracy, Formula, and Format, each comprising fine-grained criteria that reflect professional standards. The Claude family leads the benchmark and produces the most professional-looking outputs in our qualitative review, but even the strongest agents frequently fall short of professional finance standards and degrade sharply as the difficulty increases beyond a few chained calculations. This suggests that current agents are not yet able to reliably produce professional-quality spreadsheets at the level of complexity real-world workflows demand.

Problem

• LLM agents
• spreadsheet tasks
• financial workflows
• end-to-end evaluation
• professional standards

Innovation

• end-to-end spreadsheet tasks
• LLM agents
• financial workflows
• multidimensional evaluation
• professional standards