PulseAugur
EN
LIVE 09:17:50

FINCOSMOS: New Open-World Benchmark Tests AI Agents in Finance

By PulseAugur Editorial · [1 sources] ·

Researchers have introduced FINCOSMOS, a novel evaluation framework designed to assess the performance of AI agents in financial tasks within an open-world setting. Unlike existing benchmarks that constrain the environment, FINCOSMOS simulates a user with a hidden strategy, requiring agents to independently gather data from the web, conduct time-aware backtests, and deliver verifiable results. Initial testing across four coding agents and one non-tool-using model revealed significant failure rates, with human experts attributing 88.33% of these failures to the agents themselves, highlighting the challenges in real-world financial applications. AI

IMPACT Highlights critical gaps in AI agent reliability for complex, real-world tasks like finance, necessitating more robust evaluation methods.

RANK_REASON The item is a research paper introducing a new evaluation paradigm for AI agents. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

FINCOSMOS: New Open-World Benchmark Tests AI Agents in Finance

How we ranked this

Signal score
17 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
The item is a research paper introducing a new evaluation paradigm for AI agents. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
paper, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. arXiv cs.AI TIER_1 English(EN) · Yupeng Cao, Xueqing Peng, Lingfei Qian, Haohang Li, Yangyang Yu, Anke Xu, Wenbo Cao, Yan Wang, Yuechen Jiang, Yuyan Wang, Zhuohan Xie, Yankai Chen, Zehan Li, Jimin Huang, Xue Liu, Jian-Yun Nie, Sophia Ananiadou ·

    FINCOSMOS: An Open-World Evaluation Paradigm for Agents in Finance

    arXiv:2611.07464v1 Announce Type: new Abstract: Language agents increasingly work for users on the open Web, and finance is where trusting them is hardest. The data must come from the Web, the strategy lives in the user's head, and correctness depends on time, so an agent can fai…

RELATED ENTITIES

RELATED TOPICS