# BlueFin: Benchmarking LLM Agents on Financial Spreadsheets

arXiv: https://arxiv.org/abs/2605.30907 · HTML: https://mlsanity.com/paper/2605.30907/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Financial trading and forecasting agents, one of 16 directions in LLM agents and tool use.
- Builds on 18 of its 38 references that are in the map.
- First posted on arXiv 29 May 2026.

## Placed in

- LLM agents and tool use / [Financial trading and forecasting agents](https://mlsanity.com/agents/financial-trading-and-forecasting-agents/)

## Builds on (most cited first, 18 of 18)

- [Evaluating Large Language Models Trained on Code](https://mlsanity.com/paper/2107.03374/) (2021)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [Competition-Level Code Generation with AlphaCode](https://mlsanity.com/paper/2203.07814/) (2022)
- [TAT-QA: A Question Answering Benchmark on a Hybrid of Tabular and Textual Content in Finance](https://mlsanity.com/paper/2105.07624/) (2021)
- [The Stack: 3 TB of permissively licensed source code](https://mlsanity.com/paper/2211.15533/) (2022)
- [SWE-Bench Pro: Can AI Agents Solve Long-Horizon Software Engineering Tasks?](https://mlsanity.com/paper/2509.16941/) (2025)
- [GDPval: Evaluating AI Model Performance on Real-World Economically Valuable Tasks](https://mlsanity.com/paper/2510.04374/) (2025)
- [SpreadsheetBench: Towards Challenging Real World Spreadsheet Manipulation](https://mlsanity.com/paper/2406.14991/) (2024)
- [ResearchRubrics: A Benchmark of Prompts and Rubrics For Evaluating Deep Research Agents](https://mlsanity.com/paper/2511.07685/) (2025)
- [TableLLM: Enabling Tabular Data Manipulation by LLMs in Real Office Usage Scenarios](https://mlsanity.com/paper/2403.19318/) (2024)
- [MCoNaLa: A Benchmark for Code Generation from Multiple Natural Languages](https://mlsanity.com/paper/2203.08388/) (2022)
- [OfficeQA Pro: An Enterprise Benchmark for End-to-End Grounded Reasoning](https://mlsanity.com/paper/2603.08655/) (2026)
- [SheetAgent: Towards A Generalist Agent for Spreadsheet Reasoning and Manipulation via Large Language Models](https://mlsanity.com/paper/2403.03636/) (2024)
- [MiMoTable: A Multi-scale Spreadsheet Benchmark with Meta Operations for Table Reasoning](https://mlsanity.com/paper/2412.11711/) (2024)
- [How Well Does Agent Development Reflect Real-World Work?](https://mlsanity.com/paper/2603.01203/) (2026)
- [NL2Formula: Generating Spreadsheet Formulas from Natural Language Queries](https://mlsanity.com/paper/2402.14853/) (2024)
- [FinSheet-Bench: From Simple Lookups to Complex Reasoning, Where LLMs Break on Financial Spreadsheets](https://mlsanity.com/paper/2603.07316/) (2026)
- [SpreadsheetArena: Decomposing Preference in LLM Generation of Spreadsheet Workbooks](https://mlsanity.com/paper/2603.10002/) (2026)

## Built on by (most cited first, 2 of 2)

- [FinEvo-Bench: A Longitudinal Benchmark for Self-Evolving Agents in Professional Financial Workflows](https://mlsanity.com/paper/2608.06144/) (2026)
- [One Analyst Is Not Ground Truth: Grading Agent-Built Financial Models Against Observed Professional Practice](https://mlsanity.com/paper/2607.24889/) (2026)

## Abstract (arXiv)

We present BlueFin, a benchmark that tasks large language model (LLM) agents with synthesis, manipulation, and comprehension tasks over spreadsheet workbooks in the professional finance domain. Though estimates of the global population of paying users of spreadsheet software range in the hundreds of millions -- an order of magnitude more than the estimated global population of professional developers -- comparatively fewer resources have been devoted to exploring and expanding LLM capabilities in the spreadsheet domain, with fewer still dedicated to mirroring real occupational tasks encountered by those in professional finance roles. In response, we curate a set of 131 challenging, complex tasks with real-world relevance in the domain, containing 3,225 granular rubric criteria; notably, our rubric criteria and LM judge evaluations are validated by a team of expert human annotators, resulting in high-quality, granular evaluations of complex tasks that are difficult to verify programmatically but can be reliably evaluated by an LM judge agent. Our judge achieves parity with expert consensus (α=0.826) with a macro-F1 score of 0.839. Frontier LLMs demonstrate poor performance on the challenging benchmark, with the strongest LLMs achieving less than 50% average scores across tasks -- models exhibit particular weaknesses in dynamic correctness. Our contributions include a dataset of examples across three categories of spreadsheet tasks, an open source harness and agentic evaluation framework, and a characterization of existing frontier models' performance on our benchmark.