Fannie Mae and Freddie Mac just published ten years of performance data for FICO 10T and VantageScore 4.0, handing lenders and investors a working evidence base for a two score world federal regulators have not yet signed off on.
Fannie Mae and Freddie Mac just published ten-plus years of FICO 10T historical credit-score data, paired with new VantageScore 4.0 performance records. The numbers matter less than what they signal: the agencies are pre-loading the mortgage market with the evidence base for a two-score world before federal regulators officially bless it.
The new FICO 10T file lets lenders, investors, and modelers see what a decade of business at the GSEs would have looked like under the newest U.S. mortgage credit-score model. National Mortgage News reported the dataset spans loans acquired from on or about April 2013 through September 2025. The same release added VantageScore 4.0 data covering loans acquired from on or about April 2023 through September 2025.
The release is more interesting as a strategic move than as a data drop. Fannie Mae's newsroom describes the updates as advancing credit-score modernization, and Freddie Mac has already begun accepting VantageScore 4.0 on loans it purchases. The Federal Housing Finance Agency, the regulator that oversees both enterprises, has not yet finalized its multi-model credit-score framework. Its ongoing credit-scores policy page describes a transition toward letting lenders choose between competing score models.
That gap, between a regulator still deliberating and the GSEs already conditioning the market, is the mechanism behind the news. By publishing the historical data now, the enterprises give lenders, investors, and secondary-market participants something concrete to model against: ten-plus years of hypothetical FICO 10T performance, plus a shorter but more recent VantageScore 4.0 track. Pricing models, repurchase assumptions, and underwriting standards can start to harden around that evidence long before formal approval lands.
Julie May, vice president and general manager of B2B Scores at FICO, framed the release as inviting market participants to independently validate the model's strength and assess how its use can expand access for borrowers. Bob Broeksmit, president and CEO of the Mortgage Bankers Association, praised the release and urged completion of remaining reviews so both VantageScore 4.0 and FICO 10T can be made broadly available to lenders. Both reactions come from interested parties; independent lender, investor, regulator, or consumer-advocate responses have not yet appeared in the public sources available.
Stakeholder positioning aside, the data release does three things that the simpler "Fannie and Freddie publish FICO 10T data" framing leaves out. It shifts the cost of waiting onto holdouts: any lender still building pricing and servicing assumptions around older score models now has reason to run parallel scenarios, because the agencies have made the comparison cheap. It lets investors and analytics shops pre-build the score-equivalence tables that securitization documents and loan-level pricing adjustments will eventually require. And it gives the FHFA's policy team a usable external track record to point at when its own framework is published, regardless of which way the rule lands.
Freddie Mac's published guidance on credit-score models and HousingWire's coverage of the FICO 10T data both frame the release as a transition, not a finished state. Adoption volume, pricing effects, and quantified changes in borrower access are not yet established in the public record, so any forward-looking read of impact remains anticipated rather than measured.
The next watch is whether the FHFA publishes a unified credit-score framework before year-end with both FICO 10T and VantageScore 4.0 named as acceptable models, and whether any large investor or aggregator releases competing score-equivalence research that pulls against the GSEs' framing. Until that rule lands, the file the agencies just dropped functions as the default empirical anchor that lenders will be expected to underwrite against as soon as multi-model scoring is approved.