Blog

Why is my credit score a different number on every app and site?

Three free score sources, three numbers, same day. Which one is real, what makes them disagree, and how to read and track them without chasing the wrong one.

Because they are different scores. Each app and site picks a scoring model, a version of that model and a credit bureau's credit report, and each one pulls the file on its own day. Change any of the four and the number changes. None of the numbers is fake, and none of them is the one a particular lender will use. What matters is that you know what each one is, so that you compare like with like.

The rest of this post goes question by question, using one running example with made-up numbers.

The running example

Say Rae opens three free sources on the same Tuesday. Everything below is an illustration.

Source What it says it is Bureau file Number
Bank app A FICO Score Experian 688
Card issuer's site A VantageScore 4.0 TransUnion 701
Free monitoring site A VantageScore 4.0 Equifax 694

Thirteen points separate the highest from the lowest. Rae wants to know which one to believe, and whether something is wrong.

Q: Which one is the real score?

All three are real in the sense that each is a genuine output of a genuine model on a genuine report. None of them is "the" score, because there is no single master credit score. There is a family of models, and a lender picks one. The score a lender pulls for a decision may be a model and version Rae has never seen on any app.

So the better question is "which one is closest to what a lender would use?" and the honest answer is that you usually cannot know in advance. Card issuers choose among many FICO Score and VantageScore versions and their own internal models. Mortgage lenders follow rules set by Fannie Mae and Freddie Mac that have been changing since 2025. Which FICO version does a lender pull, and can you find out? and FICO Score and VantageScore compared, and why the numbers differ cover the model side in depth.

Q: What are the actual reasons the numbers differ?

Five things, in rough order of how much they matter.

1. The model. A FICO Score and a VantageScore use different formulas. According to myFICO, a general FICO Score gives about 35 percent weight to payment history and about 30 percent to amounts owed; one source lists payment history at 41 percent for VantageScore 4.0. The weightings are not directly comparable, but they show that the two do not rank inputs the same way. Two models given the same file produce two numbers.

2. The version. Within a brand there are many versions. "A FICO Score" can mean any of several, and a source that does not tell you the version is leaving out the most useful label. A newer version may use data an older one ignores, such as rent payments, but only when the rent is reported.

3. The bureau. Equifax, Experian and TransUnion each keep a file on you. The three files are similar and not identical, because lenders do not always report to all three, and when they do they may report on different days. A different file is different input. why my three credit reports do not match each other shows how that happens and what you can do about it. A credit bureau is a data company, not a scorer; the scoring models run on its file.

4. The date. Reports change when a lender reports a new balance or payment. A score pulled this morning reflects the file as of this morning. One pulled last week reflects last week. If your card issuer reported a fresh balance in between, the two numbers can differ even with the same model and bureau.

5. What the source calls a score. Some free sources show a score that is labeled as educational or as an estimate, or that uses a model no lender uses. The label is usually in small type. If a source does not say what it is, treat the number as a rough indicator at most.

In Rae's case, rows one and two already differ on three of the five reasons: model (FICO versus VantageScore), bureau (Experian versus TransUnion) and likely date. Rows two and three share a model and version but not a bureau. The 7-point gap between those two is therefore mostly a bureau-file difference, which is a useful thing to know.

Q: Is a 13-point gap normal?

There is no published tolerance, and gaps of a few dozen points between models do not by themselves signal a problem. Rae's gap is easy to explain: different model, different bureaus, possibly different dates.

A gap is worth a second look when it is very large, when a number is far from what you would expect given your history, or when a number fell sharply with no change in your behavior. Then the right move is to read the underlying report, because a wrong entry can pull a score down under every model. how to pull your free credit reports and read each section shows how to get the reports and what to look for. If something is wrong, the bureau must reinvestigate a dispute within 30 days of receiving it, a window that can extend by up to 15 days if you send more information during the 30.

Q: Will checking several apps hurt my score?

No. Checking your own score or report is a soft inquiry, which scoring models ignore. In FICO's July 2026 Harris Poll, 27 percent of adults wrongly believed checking their own score lowers it. Rae can open every app she likes. A hard inquiry follows an application for credit, which is a different event; hard and soft inquiries: which actions trigger each and what they cost sorts them.

Q: Why did my score move a few points overnight?

Because something in the report changed, or you looked at a different source. The common causes:

  • A card issuer reported a new statement balance, so utilization changed.
  • A payment posted and a balance fell, or a new charge raised it.
  • A new account or inquiry appeared.
  • An old account aged another month.
  • The source recalculated using a newer version of its model.

Small movements are routine. A few points up or down from month to month, with no change in behavior, is usually noise from reporting dates and balances. Trends over several months mean more than a single reading.

Here is a made-up tracking log for Rae on one source (the bank app, FICO, Experian), with the one change in her behavior noted.

Month Reported card balance Limit Utilization Score on this source
January $2,000 $5,000 40 percent 688
February $1,500 $5,000 30 percent 692
March $1,000 $5,000 20 percent 697

Utilization is balance ÷ limit: $2,000 ÷ $5,000 = 40 percent, $1,500 ÷ $5,000 = 30 percent, $1,000 ÷ $5,000 = 20 percent. The score rose 9 points over two months while she paid the card down. That is a trend on a single consistent source, and it is more informative than the gap between her three sources on any one day. The exact point change is not something to predict, and other factors can offset or amplify it. The credit score ranges guide shows how the bands are commonly labeled.

Q: How do I tell what I am looking at?

Look for four labels on every score you see. If a source does not show them, find them in its help pages.

  1. Brand and version. FICO or VantageScore, and which version.
  2. Bureau. Equifax, Experian or TransUnion.
  3. Date. The day the file was pulled.
  4. Range. 300 to 850 for FICO and VantageScore 4.0; another scale means another product.

Then write them down next to the number. A one-line note such as "FICO, Experian, March 12" turns a floating figure into a data point you can compare with the next one.

Q: Which one should I track?

Pick one source and use it consistently. Consistency removes the model, version and bureau differences from the comparison, so what is left is real change in your file. The source that fits best is one that updates often, shows the labels above and shows the factors behind the number. Several free sources do. Some paid services add features such as alerts and multiple bureaus, and whether the extra features are worth a fee is a separate question; free credit monitoring or a paid plan: what each really gives you sets out the trade-offs. credit monitoring is the glossary entry.

If you apply for a card or a loan, ask the lender which score it uses, if it will say. The number you track is for direction. The lender's number is the one that matters for the decision, and you might not see it.

Q: What if one number is a lot lower than I expected?

Work through these steps.

  1. Check the labels. A different model or bureau is the most common cause.
  2. Get the underlying reports. Consumers are entitled to a free credit report from each of the three bureaus through AnnualCreditReport.com, and the bureaus extended free weekly access in 2023 and made it permanent, according to the site.
  3. Read for errors. Look for accounts you do not recognize, wrong balances or limits, payments marked late that were on time, and duplicate entries.
  4. Dispute what is wrong, in writing with the bureau, and with the lender that reported it. Keep copies.
  5. If an account is not yours, treat it as possible identity theft: place a fraud alert or a credit freeze, which is free, and call the issuer using the number on the back of your card. A freeze or fraud alert and a complaint through the CFPB at consumerfinance.gov/complaint are steps to take, not shortcuts. Do not pay anyone who promises to fix your credit for a fee in advance.
  6. If the report is accurate, the lower number reflects your file. The levers, in order of speed, are lower reported balances, on-time payment every month, fewer new applications and time.

Q: What is a sensible routine, so I do not obsess?

A monthly habit is enough for nearly everyone, and it takes about ten minutes.

  1. Pick a day shortly after your main card's statement closes, since the issuer commonly reports the balance around then.
  2. Open your one chosen source and record the number with its four labels in a note or a spreadsheet.
  3. Look at the factors the source shows, if any: utilization, payment history, age of accounts, inquiries. Note anything that changed.
  4. Do not act on a single reading. A move of a few points either way is normal. Look for a direction over three months.
  5. Once or twice a year, read the full reports from all three bureaus, because the score is only a summary and an error can hide in the detail.

The routine has a second purpose. It gives you a baseline so that if something odd shows up, such as a sudden drop or a new account you do not recognize, you notice it within a month, not a year. Early notice makes an identity-theft problem easier to contain. The causes of an unexpected drop are worth working through one at a time.

Q: What should I watch for when a site offers a free score?

A few plain precautions.

  • Start from a source you already trust, such as your bank, your card issuer or a bureau, or go to AnnualCreditReport.com for reports. Be careful with a link from an email or text, and with a caller who says your score needs attention. A message that pushes you to act fast, share a Social Security number or pay a fee to "fix" a score is a warning sign.
  • Read what you are signing up for. Some free score offers lead into a paid trial. Check the terms, the billing date and how to cancel before you enter payment details.
  • Do not pay anyone to remove accurate items. Accurate negative items stay for the period the law allows, which for most items is seven years. Disputing an error is free and something you can do yourself.
  • Use the number on the back of your card for questions about a specific account, and the CFPB complaint process if an issuer or a bureau will not fix a documented error.

A score app is a tool for looking at your file. The file itself, and what you do with it, is what moves the number.

Q: Does the average score tell me anything?

A little, as context. FICO reported an average U.S. FICO Score of 714 in its Spring 2026 report and said in August 2026 that the average held at 714 from October 2025 to April 2026. Experian's 2025 review cited an average of 713. Those are population averages from the publishers, and they say nothing about how a given file will score under a lender's model. Rae's three numbers, 688 to 701, sit a little below those averages, and that is a fact about where she stands among consumers, not a prediction of any decision.

Q: Is a higher number from one source a reason to use it?

No. A higher number from a different model does not mean the same thing as a higher number from the one you tracked. It also does not mean a lender will see it. It is tempting to quote the highest number in a conversation with a lender or landlord. They will pull their own.

Key takeaways

  • The three numbers differ because each source picks its own model, version, bureau file and pull date, and none is the single master score; no lender is obliged to use any of them.
  • In Rae's example, the VantageScore sources differed by 7 points on a bureau change alone, and the FICO source differed from them on model and bureau together.
  • Checking your own score is a soft inquiry and does not lower it, so check as many sources as you like.
  • Track one source consistently and note its brand, version, bureau and date; Rae's 9-point rise while utilization fell from 40 to 20 percent is a clean trend on one source.
  • For a surprising number, read the underlying reports from AnnualCreditReport.com and dispute errors; a bureau must reinvestigate within 30 days of receiving a dispute.
  • When applying, ask the lender which score it uses; the number you track is for direction.

Terms used in this article

Related tools

Latest on this topic