Credit & Credit ScoresAdvanced6 min read

How lenders actually price you: tiers, rate sheets, and the cutoffs that matter

Your rate isn't calculated — it's looked up. Inside risk-based pricing: the tier tables, mortgage pricing grids, and threshold scores where two points of FICO cost thousands.

Borrowers imagine their interest rate is computed from their file with actuarial precision — a 703 score earning a slightly better rate than a 701. The reality is cruder and more gameable: lenders price from rate sheets, tables that map score ranges and loan characteristics to rates. Within a bucket, everyone pays the same. At the boundary between buckets, a two-point score difference repriced the entire loan. Understanding where those boundaries sit — and that they differ by product — turns credit improvement from a vague virtue into a targeting exercise.

Risk-based pricing in one paragraph

Lenders price to expected loss: the probability a borrower like you defaults, times the loss when it happens, plus funding costs and margin. Scores exist to sort borrowers into probability bands, and because building, testing, and regulating a smooth pricing curve is expensive, the industry quantizes the curve into tiers. A rate sheet is that quantization printed out: rows of score ranges, columns of terms or loan-to-value ratios, a rate in each cell. Loan officers don't negotiate your risk — they find your cell.

Auto lending: the tier system in its purest form

Tier	Score band	Typical used-car APR
Super prime	781–850	~7%
Prime	661–780	~9–10%
Near prime	601–660	~14%
Subprime	501–600	~19%
Deep subprime	300–500	~22%+

Typical auto-lending tiers and used-car APRs (representative estimates; sheets vary by lender)

The $4,700 gap between 659 and 663

Two buyers finance the same $28,000 used car over 72 months. At 659 — near prime on most sheets — the rate is about 14%: a $577 payment and roughly $13,500 in total interest. At 663, just across the prime line, the sheet says about 10%: $519 a month and roughly $8,800 in interest. Four points of score, $4,700 of interest, because both borrowers were priced as their tier, not as themselves. This is why paying a $40 collection or letting a card balance report low the month before a car purchase can be the highest-ROI financial move available to a 650s borrower — and why buying a car the month your score sits at 655 instead of waiting one statement cycle is so expensive.

Mortgages: a grid, not a ladder

Conventional mortgage pricing runs on loan-level price adjustments — a published grid of fees crossing score bands (typically in 20-point steps: 640–659, 660–679, 680–699, 700–719, 720–739, 740–759, 760–779, 780+) against loan-to-value ratios. Each cell adds or subtracts cost, usually absorbed into the rate. Two consequences matter. First, the boundaries are knife-edged: 739 vs. 740 changes your cell, which commonly moves the rate an eighth to a quarter point — worth several thousand dollars over a 30-year loan. Second, the grid means score and down payment interact: a bigger down payment can partially offset a middling score and vice versa, so the cheapest path to a better cell isn't always the credit side. Mortgage lenders also pull all three scores and use the middle one (the lower of two middles for couples), so the weakest bureau of the weaker borrower often controls the cell — a specific, checkable thing to optimize before applying.

Cards and personal loans: tiers in disguise

Credit card offers advertise APR ranges — '19.24%–28.24% variable' — and the range is the tier system showing through: your approval lands you at one of a handful of price points within it, set by score band and income. Personal-loan marketplaces work the same way, which is why prequalification (a soft pull returning your actual offer) is the single best pricing tool consumers have: it reveals your cell at multiple lenders without inquiry damage. Two lenders can place the same borrower two tiers apart, because each calibrates its own sheet to its own loss history — the spread between the best and worst legitimate offer on the same profile is routinely five-plus points of APR.

Roughly where the big pricing cliffs sit (varies by lender and product)

580 — FHA minimum for 3.5% downMortgage access

620 — conventional mortgage floorAccess + worst grid cells

661 — prime auto tier beginsAuto APR drops hard

700 — mid-grid mortgage stepBetter LLPA cell

740 — historically strong mortgage tierMost of the grid benefit

781 — super prime autoBest sheets everywhere

Find your product's boundaries before optimizing: 740 matters for a mortgage, 661 for a car, and neither matters much for a card you'll pay in full.
If you're within 10–15 points below a boundary, delay the application one or two statement cycles and force utilization low — the cheapest tier-jump in finance.
Ask lenders directly: 'What score bands does your pricing use, and which bureau and model do you pull?' Loan officers will usually tell you, because it helps them close.
Prequalify at three or more lenders for any auto or personal loan — you're shopping sheets, not just rates, and sheets disagree.
Remember rate isn't the only tiered variable: limits, down-payment requirements, and mortgage-insurance pricing all step at similar boundaries.

The score you watch isn't the score they price

Free score apps typically show VantageScore or a generic FICO 8 from one bureau. Mortgage lenders still predominantly price on the classic FICO 2/4/5 bureau models, auto lenders often use FICO Auto Score variants, and card issuers use bankcard-tuned versions — each can differ from your app's number by 20 to 40 points, in either direction. Before a major application, check the score family the lender actually uses (myFICO sells the mortgage and auto versions; some lenders disclose theirs at prequalification). Optimizing the wrong ruler feels productive and moves nothing.

Timing beats grinding near a boundary

Scores wobble 5–15 points month to month with reported balances. If you're at 736 chasing a 740 grid cell, you often don't need months of credit work — you need the right snapshot: pay all card balances to near zero before the statements cut, let them report, then apply during that reporting window. Lenders price the photograph, not the film.

The bottom line

Risk-based pricing sounds continuous but ships in steps: tiers on a sheet, cells on a grid, bands behind an APR range. That structure is exploitable in your favor — find the boundary your product uses, check the score model that lender prices on, time your snapshot, and shop multiple sheets. The difference between a 659 and a 663 isn't four points of creditworthiness; it's whichever side of a line a committee drew. Make sure you're standing on the cheap side when the photo gets taken.

Check your understanding

1 of 4

Two buyers finance the same $28,000 car — one at a 659 score, one at 663. Why did four points cost the first buyer roughly $4,700?

Not quite — try again.

The Worth letter

Get smarter about money every week

One email, no spam — practical guides and Worth updates. Unsubscribe anytime.

Put this into practice

Worth tracks your accounts, budgets, and goals — so the concepts in this article aren't just theory.

Start free trial