HousingWire: Tidalwave AI tops general models in mortgage underwriting accuracy benchmark
A Columbia University and Tidalwave benchmark study tested 90 underwriting questions across 10 synthetic loan files. Tidalwave’s SOLO scored 84% accuracy compared to 71% for Claude 4.5, including 95% to 42% on yes-or-no compliance checks, while Claude led on account verification.
Read the full post
