跳到正文
arXiv cs.AI· Elaine Lau, Antoine Pangas, Alexandre Finot, Collin Schweiker, Amir Bouzidi, Hui Wen Goh, Andrew Peter Martin, Jonas Mueller·· 56 分钟前AI 评分55

ATLAS-Finance 基准发布:100 项投行任务测试 AI 智能体,最强模型通过率不足 11%

ATLAS-Finance: Benchmarking AI Agents in Realistic Banking Environments

AI 导读

论文发布 ATLAS-Finance 基准,包含跨 13 家模拟公司的 100 项端到端投资银行任务,由资深银行家基于真实交易工作开发,人类需 15-30 小时完成。智能体需搜索邮件、聊天、日历、笔记和文件来收集信息、调和冲突并构建复杂金融模型,交付物由专家评分标准(中位数 124.5 项准则)评估。结果显示当前前沿模型中表现最好的 Claude Opus 和 Fable 通过率不足 11%。

正文 · 原文

View PDF HTML (experimental)

Abstract:Professional work involves under-specified assignments, where relevant information is spread across workplace applications and may conflict or change over time. To evaluate how well frontier agents can perform economically valuable work in such settings, we introduce ATLAS-Finance, a benchmark of 100 end-to-end investment-banking tasks across 13 simulated firms. As in real banking workflows, agents must search email, chat, calendars, notes, and files to identify requirements, gather information, reconcile conflicts/updates, and then build complex financial models. Tasks were developed by experienced bankers based on their real deal work and take them 15-30 hours to complete. Agent deliverables are evaluated with expert-authored rubrics (median: 124.5 criteria) covering financial rationale and results, spreadsheet construction, completeness, and formatting. Despite completing substantial portions of the requested work, current agents rarely produce correct deliverables. Across today's frontier models, the best ones (Claude Opus and Fable) pass fewer than 11% of tasks.
Subjects: Artificial Intelligence (cs.AI)
Cite as: arXiv:2611.22085 [cs.AI]
  (or arXiv:2611.22085v1 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2611.22085

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Elaine Lau [view email]
[v1] Tue, 17 Nov 2026 02:12:40 UTC (336 KB)

来源:arXiv cs.AI · arxiv.org