Answers are grounded in this paper and cite their sources.
Probability calibration, directional discrimination, and portfolio performance measure different properties of financial news models. Evaluating TypeSafe Jev (jev-1.13.0), GPT-5.6-sol, DeepSeek-V4.1-flash, and verbalized Qwen3-8B across 150,623 news articles, 103,279 ticker-days, and 361 equities over 2019–2023, we document their empirical separation across the evaluated models and samples. Ranking associations are strongest in the close-to-open interval, making signal availability and attainable entry prices central to economic interpretation.
No takes yet. Share an insight, caveat, or question.
Luo et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: