VantageScore says its latest performance assessment confirms that VantageScore 4.0 is better at predicting mortgage risk than earlier VantageScore models and other benchmark credit scores, strengthening the company’s case as lenders begin working with multiple approved scoring models.
The company released its eighth annual Model Performance Assessment on Tuesday, covering consumer-credit performance from June 2023 through June 2025. VantageScore said the model outperformed prior versions and benchmark credit-scoring models in originations and account management, including across mortgage categories.
VantageScore attributed the results partly to the model’s use of trended credit data, which examines how consumers manage balances and payments over time instead of relying only on a snapshot of their credit profile.
“Mortgage credit scoring has entered a new era where precision, consistency and transparency are no longer optional. They are mandatory,” said Dr. Andrada Pacheco, chief data scientist at VantageScore. “VantageScore 4.0 continues to demonstrate how advanced analytics and trended credit data enable lenders to make more informed decisions while expanding responsible access to credit for millions of creditworthy consumers.”
VantageScore said its model uses four times as much data as legacy mortgage credit scores, providing what it described as a more predictive view of borrower creditworthiness.
The company also said VantageScore 4.0 produced the most consistent results among its scoring models across Equifax, Experian, and TransUnion. VantageScore attributed that consistency to its proprietary attribute-leveling process, which is intended to reduce score differences resulting from variations in how the three nationwide credit bureaus structure consumer-credit information.
That could be particularly relevant in mortgage lending, where lenders continue to obtain credit information from all three bureaus. Greater consistency does not mean the three scores will be identical, but it could reduce some of the variation originators encounter when evaluating a borrower’s qualifying score.
Competing Research Reaches A Different Result
VantageScore’s findings enter a market where the performance of newer credit scores remains contested.
An analysis by actuarial firm Milliman found in May that FICO Score 10T outperformed VantageScore 4.0 across major mortgage categories, including conforming and Federal Housing Administration loans. The study examined nearly 20 million mortgage tradelines and was commissioned by FICO.
After Fannie Mae and Freddie Mac released expanded historical loan-level credit-score data July 1, Milliman conducted another FICO-commissioned analysis. FICO said that analysis found FICO 10T outperformed VantageScore 4.0 across the K-S statistic, bottom-decile lift, and Gini coefficient, three measures used to evaluate how accurately a score separates borrowers by default risk.
The public GSE data allow lenders, investors, and researchers to conduct their own comparisons. However, the Milliman work was commissioned by FICO, just as the latest assessment was performed by VantageScore on its own model.
The competing findings do not independently settle which model performs best across every mortgage population. They do reinforce the need for lenders to examine the methodology, borrower sample, performance period, and definition of default behind each claim.
What It Means
For originators, the practical issue is whether a borrower’s score and loan execution differ under the models available through a lender.
A borrower who misses an eligibility or pricing threshold under Classic FICO could receive a different score under VantageScore 4.0. That difference may affect loan-level price adjustments, mortgage insurance, available terms, or whether the loan receives an eligible automated underwriting result.
It does not guarantee approval. Borrowers must still satisfy income, asset, debt-to-income, collateral, and other underwriting requirements.
VantageScore has argued that its broader scoring criteria could be particularly useful for consumers with thin, inactive, or nontraditional credit histories. In July, the company estimated that VantageScore 4.0 could identify more than 5 million additional creditworthy mortgage borrowers, representing as much as $1 trillion in potential origination volume.
Those figures also came from VantageScore research and represent an estimated addressable market, not closed-loan volume.
For brokers, the latest assessment adds another question to lender selection: whether a wholesale partner supports VantageScore 4.0, for which products, and whether evaluating a borrower under the alternative model produces a stronger defensible execution.
