Sonto Research / Evals

Release
Version 1
Frozen

Frontier Financial Judgement: Can agents tell what might move a stock?

A benchmark of whether research agents can distinguish genuinely new, valuation-relevant financial information from realistic noise.

82

Cases

14

Models

656

Supplied items

246

Expert-assigned labels

Important

All events below are synthetic and do not reflect any real-world developments. We have removed company names from key parts of the text, titles, and expert rationale.

Loading the selected benchmark case.