For full context here is a related question you posed - How can I understand historical changes in my credit score?
One of the points I made there which, for the sake of posterity, I'll repeat here is that you can't reconcile the scores calculated using 2 different models. Because:
- Algorithm is different
- Service providers are different and operate at their cadence
Short answer: Don't stress about it
Not just the model but when they poll the credit card data and what the data shows at that time is arbitrary (but at a cadence).
For ex, credit cards by themselves will not post their data to the scoring application. The only way a scoring application (service provided for free by credit cards, credit monitoring apps like Credit Karma) can do the scoring is to decide when to poll all the accounts you have. On that date, whatever credit info (like utilization, age of accounts, new accounts opened, timely payments, loan, mortgage) is taken and a score is calculated. This may fall on a date that is different from the statement date or payment date of your cards.
Which means, if they had waited until you eventually paid the bill (even on time), they'd have seen close to 0% utilization. But then some other card having different cycles may have higher numbers. So all you can do is to use these tools to generally monitor the health, get an indicator of the scores to have an idea, correlated score changes with activities and report incorrect activities (not scoring cycles or reconciliation).
And when you actually apply for a new loan (auto/home) the score by a particular model maybe irrelevant because often, they take more factors into account like income, expenses, collateral, payment history (without assigning numbers). So, don't sweat it. Look for discrepancy in the activities and worry about fixing those with the bureaus/lenders.
Personally, I have 2 scores differing by 100 and some differing by 20s at any given time. Looking further it looks like some card info (as seen by one of the apps) is delayed by even 2 months. This is, likely, because the last time they polled there may have been an error, or the credit card returned the same info, etc. (Software wise). But the info returned by all of them are the same wrt the factors (utilization, balances, timeliness, age, etc)