Researchers have introduced Earnings25, a new benchmark designed to evaluate automatic speech recognition (ASR) systems specifically on financial earnings calls. This benchmark includes two datasets: a 498-hour set of full earnings calls from S&P 500 companies in Q4 2025, and a 46-hour segmented set from U.S. earnings calls in 2025. Earnings25 offers aligned transcripts and metadata such as speaker roles and industry labels, allowing for more nuanced evaluations beyond simple word error rate. AI
IMPACT This benchmark could drive improvements in ASR accuracy for financial sector applications.
RANK_REASON The cluster describes a new academic benchmark for evaluating ASR systems. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Hugging Face Daily Papers →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
