Metraipolis › archive › dc_bd80889faad1
FinExam-10K: When Retrieval Helps Financial Reasoning? — arXiv:2608.28155
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.28155
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T10:56:35.310Z
- SHA-256 of the archived text
a2c5d5fe35b0ead0bd6a1eb63e6e19f56d5cfa6d0d641220e22b73b28a8bdfa5- Stable URL of this page
https://metraipolis.com/documento/dc_bd80889faad1
How to cite
FinExam-10K: When Retrieval Helps Financial Reasoning? — arXiv:2608.28155 — arXiv (en), https://arxiv.org/abs/2608.28155 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bd80889faad1 (Metraipolis archive dc_bd80889faad1, sha256 a2c5d5fe35b0ead0bd6a1eb63e6e19f56d5cfa6d0d641220e22b73b28a8bdfa5)
BibTeX
@misc{dcbd80889faad1,
title = {FinExam-10K: When Retrieval Helps Financial Reasoning? — arXiv:2608.28155},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.28155},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bd80889faad1; sha256 a2c5d5fe35b0ead0bd6a1eb63e6e19f56d5cfa6d0d641220e22b73b28a8bdfa5},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.28155. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Professional financial examinations require models to combine domain knowledge, calculation, and judgment, yet no benchmark covers the full CFA and FRM structure under one protocol. We introduce FinExam-10K, to our knowledge the largest reported English benchmark for this setting, with 10,198 expert-reannotated questions spanning CFA Levels I-III and FRM Parts I-II. We release 5,110 questions and sequester 5,088 for a quarterly maintained leaderboard. To separate coverage from local answerability, we report a 10,198-item Full-Coverage Track and a 7,625-item Context-Complete Reasoning Track, which is the primary basis for claims about reasoning from the supplied record. Across 17 models, the best accuracy is 85.29% overall. On the frozen Hard band, the best score is 34.68% on the Full-Coverage Track and 54.57% on the 372 context-complete items. All 17 models share 47 context-complete failures. Function-RAG and FunctionGraph-RAG rescue hundreds of errors but also overturn many correct answers, producing little or negative net gain. A gate trained only on public data decides from the question and initial response when FunctionGraph-RAG should run. On the 5,088 held-out items, the gate invokes FunctionGraph-RAG for 7.9% of questions and improves accuracy from 70.83% to 71.23% (p = .0446). Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-31 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2608.28155 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T10:47:33.131Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-31?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).