Overview
FinMMEval 2026 Task 2 evaluates short-answer financial question answering over multilingual evidence. Each final-test item pairs an English question with financial statements and news in Chinese, Japanese, Spanish, and Greek. Participating systems submit one concise answer per item in JSONL format.
Test Set Composition
The final test set contains 256 items, evenly split between easy and expert tiers; each tier has four question templates instantiated over 32 company-report groups. Gold answers were withheld during submission, and systems were ranked by macro-averaged item-level ROUGE-1 F1 against organizer-held reference answers.
Ranking and System Performance
The final leaderboard includes 12 ranked submissions, with the strongest systems closely clustered, the top four separated by less than one percentage point in ROUGE-1 F1. The submitted system papers document retrieval-augmented generation, cross-lingual evidence handling, structured prompting, answer compression, and validation strategies.
Blogger's Review: The design of FinMMEval 2026 emphasizes the processing of financial information in multilingual contexts, showcasing the potential of natural language processing in the finance domain. The diverse strategies of participating systems provide valuable insights for future research in question-answering systems.