New Earnings25 benchmark evaluates ASR for financial earnings calls

Researchers have introduced Earnings25, a new benchmark designed to evaluate automatic speech recognition (ASR) systems specifically for financial earnings calls. This benchmark includes a substantial dataset of nearly 500 hours of S&P 500 earnings calls from Q4 2025, along with a segmented set of 46 hours representing various industries. Earnings25 offers aligned transcripts and metadata such as speaker roles and industry labels to facilitate more nuanced evaluations beyond simple word error rate. AI

IMPACT Enables more accurate evaluation of ASR systems in the finance sector, potentially improving tools for financial analysis and reporting.

RANK_REASON The item describes a new academic benchmark for evaluating AI systems, published on arXiv. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• CatalyzeX
• DagsHub
• Earnings25
• Gotit.pub
• Hugging Face
• Parakeet-TDT
• ScienceCast
• S&P 500
• Whisper

• paper
• other

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