Rate and loan-limit figures in this article are current as of September 2026 and reflect Mortgage Bankers Association, Freddie Mac, and Federal Housing Finance Agency data; mortgage pricing moves weekly and individual quotes will differ. This is educational analysis, not lending advice.
TL;DR — Quick Verdict
- The latest MBA survey with a fully disclosed rate-and-points breakdown, for the week ending September 4, 2026, showed 6.74% jumbo versus 6.85% conforming, with 0.63 points versus 0.67 points respectively — an 11-basis-point advantage for jumbo.
- MBA’s dashboard has since shown the overall 30-year fixed rate climbing further, to 6.97% for the week ending September 11, 2026, but a jumbo/conforming split for that week could not be independently sourced for this update; the loan-level comparisons below use the September 4 reading.
- The 2026 baseline conforming loan limit is $832,750, up $26,250 from 2025. The high-cost-area ceiling is $1,249,125.
- On a $900,000 loan, using the September 4 MBA rate-and-points data, jumbo financing costs about $66 less per month and about $23,800 less in 30-year interest than a hypothetical conforming-priced loan at 6.85%. The 0.04-point difference adds another $360 in upfront cost savings.
- The rate gap is small enough that lender selection matters more than the loan category. Compare at least three quotes on the same day, and price a jumbo against a conforming-first-lien-plus-second-lien structure if you are close to your county limit.
Conventional wisdom says jumbo mortgages cost more. The latest MBA data still says otherwise. For the week ending September 4, 2026, the Mortgage Bankers Association recorded an average contract interest rate of 6.74% on 30-year fixed-rate jumbo loans against 6.85% on conforming loans of $832,750 or less — an eleven-basis-point advantage for the larger loan. Jumbo borrowers also paid fewer points: 0.63 versus 0.67, including origination fees, at 80% loan-to-value.
The inversion is not permanent. MBA data shows the jumbo-conforming spread moving in both directions over the past several weeks. Jumbo was 0.09 percentage point cheaper than conforming in the week ending July 31 and again in the week ending August 7, before narrowing to 0.05 point in the week ending August 21 and to 0.03 point by August 28. It then widened sharply to 0.11 point in the week ending September 4 — the widest gap in this six-week window. The latest rate reading therefore supports the same basic conclusion as the August version of this article, but the advantage has widened again.
This article breaks down the current spread with MBA and Freddie Mac survey data, models the 30-year cost difference on a $900,000 loan, explains why jumbo pricing behaves the way it does, and identifies the borrower profiles where each product genuinely wins. Lenders including Chase, Bank of America, Rocket Mortgage, and regional portfolio lenders price these two products from different funding sources — which is exactly why the gap moves.
Current Jumbo and Conforming Rate Data
Two national surveys track these products weekly. The MBA’s Weekly Mortgage Applications Survey reports separate contract rates for conforming and jumbo balances, making it the most useful widely published source for reading the spread directly. Freddie Mac’s Primary Mortgage Market Survey covers conforming loans only — its latest verified reading available in the current research set was 6.76% for the week of September 10, 2026, up from 6.71% the week before, while the MBA’s latest published dashboard reading shows an overall 30-year fixed rate of 6.97% for the week ending September 11, 2026, without a confirmed jumbo/conforming split for that specific week.
The latest MBA rate-and-points release that can be paired cleanly for cost calculations is the week ending September 4, 2026. MBA reported 6.85% conforming and 6.74% jumbo, with 0.67 and 0.63 points respectively at 80% LTV. The MBA dashboard subsequently displayed a higher composite 30-year reading for the week ending September 11, but a newer fully itemized rate-and-points release for that week was not independently verified for this update, so the cost model in this article uses the September 4 figures throughout.
Here is how the spread has moved across recent survey weeks. Direction has reversed repeatedly.
Source: Mortgage Bankers Association Weekly Mortgage Applications Survey, 80% LTV, 30-year fixed-rate contract rates. Points include origination fee. Positive spread means jumbo costs more. The August 21, 2026 row is derived from the prior-week comparison published in MBA’s August 28 release rather than from a separately retrieved August 21 press release. The September 4 reading is the latest fully itemized rate-and-points observation used for the article’s cost model.
Across these five survey weeks, the jumbo-conforming spread ranged from −0.03% (August 28) to −0.11% (September 4), narrowing through most of August before widening again in the most recent reading. That is a materially different picture from the old assumption that jumbo loans carry a fixed premium. For borrowers tracking the broader conforming market, current 30-year fixed rate data provides the baseline against which both products price.
What Actually Determines the Jumbo Premium
Two entirely separate funding machines produce these rates, and that architecture explains why the spread can move so quickly.
A conforming loan gets sold to Fannie Mae or Freddie Mac, packaged into a mortgage-backed security, and priced off investor demand in the MBS market. The originating lender collects a fee and moves on. Because the government-sponsored enterprises buy at a published price, conforming rates cluster relatively tightly across lenders — the raw funding economics are more standardized. That price is then adjusted by loan-level price adjustments, a cumulative fee grid set under the GSE pricing framework that can surcharge lower credit scores and higher loan-to-value ratios.
Jumbo loans have no such agency buyer. A lender either holds the loan on its own balance sheet or sells it into a much thinner private-label securitization market. When a large bank wants deposit relationships with high-income households, it can price jumbo mortgages as a relationship product, sometimes below what a simple MBS comparison would suggest. That is one mechanism behind today’s inverted spread.
Consider a specific case. A software director in Bellevue, Washington buys at $1,125,000 with 20% down, needing a $900,000 loan. King County has a high-cost conforming limit above the national baseline, but whether a specific $900,000 loan is jumbo depends on the property’s county-level limit. Her national bank quotes 6.74%. A credit union quotes 6.99% because it lacks the same balance-sheet capacity. Same borrower, same week, 25 basis points apart — more than double the current national category spread. The differences between bank, credit union, and online lender pricing can dwarf the jumbo-conforming gap, and credit score tier effects on pricing can move a quote further still.
The 2026 Conforming Loan Limits That Define the Boundary
Whether a loan is jumbo is a geographic question, not a wealth question. The FHFA sets limits annually under the Housing and Economic Recovery Act, indexed to changes in home values. Home values rose 3.26% between the third quarters of 2024 and 2025, so the 2026 baseline rose by exactly that percentage.
Source: Federal Housing Finance Agency, Conforming Loan Limit Values for 2026, announced November 25, 2025. FHFA conforming loan limit values. County-level limits vary; verify your county on the FHFA limit map.
Limits rose in all but 32 counties or county equivalents for 2026. If you were quoted as a jumbo borrower in late 2025 on a balance between $806,500 and $832,750, that same loan is conforming today in counties where the baseline limit applies.
Jumbo vs Conforming: Which Is Better for a $900,000 Loan?
Run the numbers on a borrower who can go either way — someone in a baseline county buying with enough cash flexibility to either take a $900,000 jumbo or put more down and land at $832,750 conforming.
Scenario A — single $900,000 jumbo at 6.74%. Principal and interest come to $5,832 per month. Total interest over 30 years: approximately $1,199,470. Using the latest fully disclosed MBA points figure of 0.63, points add $5,670 at closing.
Scenario B — $832,750 conforming at 6.85% plus $67,250 cash. Monthly principal and interest: $5,898. Total interest over 30 years: approximately $1,223,280. Using the latest fully disclosed MBA points figure of 0.67, points add approximately $6,030. But this requires an additional $67,250 out of pocket at closing.
Scenario B looks cheaper in total interest because it borrows $67,250 less. The correct comparison holds the borrowed amount constant.
Original calculation by Real Cost Report using standard amortization, based on the MBA contract rates and points for the week ending September 4, 2026. The second row is modeled, not offered — conforming pricing does not apply above the applicable county loan limit.
Verdict
At the latest fully disclosed MBA rate-and-points pricing, the jumbo loan wins by roughly $24,170 across points and 30-year interest on the same $900,000 borrowed amount. The monthly payment difference is only about $66, so the decision should not turn on the national category spread alone. Draining $67,250 in reserves to squeeze under a conforming limit can be counterproductive when a jumbo lender requires substantial post-closing liquidity. Price both structures, but keep the reserve requirement in the comparison.
What Most Borrowers Get Wrong About Jumbo Pricing
Mistake 1: Assuming jumbo always costs more. The consequence is anchoring to a bad quote and never shopping the jumbo market. MBA data shows jumbo pricing below conforming in every one of the last several survey weeks, although the size of the advantage has changed materially from week to week. Correct action: pull the current MBA survey before you accept any lender’s framing of a “jumbo premium.”
Mistake 2: Comparing rate without comparing points. A 6.50% jumbo quote at 1.25 points costs more than a 6.75% quote at zero points for anyone who moves within a relatively short period. In the latest fully itemized MBA week, conforming borrowers paid 0.04 points more than jumbo borrowers on average — worth $360 on a $900,000 loan. Correct action: compare on annual percentage rate and total five-year cost, using an APR-based lender comparison rather than headline rate.
Mistake 3: Making a large extra down payment purely to reach conforming territory. This drains reserves that jumbo underwriters may require and converts liquid capital into illiquid equity for a spread that can be negative. Correct action: model both structures before committing cash, and treat reserve depletion as a cost.
Mistake 4: Ignoring the piggyback alternative. Borrowers $50,000 to $150,000 above their limit often skip the conforming-first-lien-plus-second-lien structure entirely. When jumbo pricing runs above conforming, that structure can win — when jumbo runs below, as in the latest MBA observations, it may lose. Correct action: price it once, both ways.
Mistake 5: Locking without understanding expiration risk. Jumbo files can carry heavier documentation and take longer to clear underwriting, which makes a 30-day lock a meaningful consideration. Extensions are not free. Review rate lock duration and extension costs before choosing a lock period on a jumbo file.
Who Should Choose Jumbo, and Who Should Stay Conforming
Structure follows profile, not preference. Four conditional rules cover many situations, but lender overlays can change the answer.
Take the jumbo if your loan exceeds the county limit by more than $150,000. Below that gap a piggyback structure is at least worth pricing; above it, the second lien gets large enough that its rate — typically well above first-lien pricing — can erase any advantage.
Stay conforming if your credit profile does not meet available jumbo underwriting standards. Jumbo lenders commonly impose higher credit-score, reserve, debt-to-income, and documentation requirements than standard conforming programs, but there is no single agency-set jumbo minimum. Borrowers who cannot qualify for a competitive jumbo quote should also weigh FHA versus conventional total cost, where applicable.
Take the jumbo if you have enough post-closing reserves to satisfy the lender. Reserve requirements are one of the hardest jumbo hurdles for otherwise-qualified buyers, and requirements vary substantially by lender, loan size, property type, and borrower profile.
Consider a jumbo adjustable-rate mortgage if your holding period is under seven years. Jumbo ARM pricing can undercut jumbo fixed pricing because portfolio lenders have different funding and asset-liability considerations. The math depends entirely on your timeline — work through the ARM versus fixed break-even calculation before assuming a fixed rate is safer in cost terms.
One structural note for anyone timing a purchase: mortgage pricing tracks the 10-year Treasury yield far more closely than Federal Reserve policy moves, which is why mortgage rates and Fed decisions diverge so often. Jumbo pricing adds a second variable on top — bank funding costs and balance-sheet appetite — making it the more volatile of the two products week to week.
The Fed’s September 16, 2026 decision confirms the direction the July vote pointed toward. The FOMC voted 12–0 to raise the federal funds target range by a quarter point to 3.75%–4.00% — its first increase since July 2023 — and its updated projections point to a possible additional quarter-point hike before year-end. Much of this move was already priced into Treasury yields and mortgage rates by early September, which is part of why jumbo and conforming rates kept climbing through the survey weeks in this update even before the meeting took place. The decision removes some uncertainty, but it doesn’t guarantee mortgage rates now fall, since long-term rates track the 10-year Treasury more than the fed funds rate itself.
Frequently Asked Questions
Why are jumbo rates lower than conforming rates in 2026?
Jumbo loans generally stay on lender balance sheets or move through private-label channels rather than being sold directly to Fannie Mae or Freddie Mac. Large banks can also compete aggressively for high-income customers and deposit relationships. The result is that jumbo pricing can sometimes undercut conforming pricing. In the latest fully itemized MBA rate reading, for the week ending September 4, 2026, jumbo was 6.74% versus 6.85% for conforming.
What is the conforming loan limit in my county for 2026?
The national baseline is $832,750 for one-unit properties in most U.S. counties, rising to a ceiling of $1,249,125 in high-cost areas. Alaska, Hawaii, Guam, and the U.S. Virgin Islands have special limits, including a 2026 ceiling of $1,873,675. Limits rose in all but 32 counties or county equivalents for 2026. The FHFA publishes county-level limits — check yours directly rather than assuming the baseline applies.
Does a jumbo loan require a bigger down payment?
Often yes, but there is no single jumbo down-payment rule. Some lenders offer 10% down or other structures for strong borrowers, while second homes and investment properties can require substantially more. Requirements vary by lender because no federal agency sets a universal jumbo underwriting standard. Get quotes from at least three institutions and compare the required reserves as well as the down payment.
Should I buy points on a jumbo loan?
It depends on your break-even horizon. In the MBA survey for the week ending September 4, 2026, jumbo borrowers averaged 0.63 points — $5,670 on a $900,000 loan. Because jumbo balances are large, each point represents a large absolute dollar amount. That can shorten the break-even period if the rate reduction is meaningful, but it also means an early sale or refinance can forfeit more upfront cash. Model your specific holding period first.
How We Researched This Article
Rate figures come from the Mortgage Bankers Association Weekly Mortgage Applications Survey, which reports average contract interest rates and points separately for 30-year fixed conforming balances ($832,750 or less) and 30-year fixed jumbo balances (greater than $832,750), both at 80% loan-to-value. The latest fully itemized observation used for the article’s rate-and-points calculations is the week ending September 4, 2026: 6.85% conforming and 6.74% jumbo, with 0.67 and 0.63 points respectively. MBA’s dashboard subsequently displayed a higher composite 30-year reading for the week ending September 11, 2026 (6.97%), but a fully itemized jumbo/conforming breakdown for that week could not be independently verified for this update; the September 4 figures are retained as the latest confirmed baseline.
For historical context, we used MBA observations for the weeks ending July 31, August 7, August 21, August 28, and September 4, 2026. The August 21 reading is derived from the prior-week comparison published in MBA’s August 28 release, since a standalone August 21 press release was not directly retrieved for this update. These show that the jumbo-conforming spread is not fixed: jumbo was below conforming by 9 basis points on July 31, 9 basis points on August 7, 5 basis points on August 21, 3 basis points on August 28, and 11 basis points on September 4.
Freddie Mac’s Primary Mortgage Market Survey supplied the latest independently verified conforming benchmark available in the current research set: 6.76% for the week of September 10, 2026, compared with 6.71% the prior week and 6.66% two weeks earlier. Freddie Mac’s survey covers conventional conforming purchase loans and does not track jumbo pricing.
Loan limit values come directly from the Federal Housing Finance Agency announcement of November 25, 2025. The 2026 national baseline is $832,750, up $26,250 from 2025. The national high-cost ceiling is $1,249,125. Alaska, Hawaii, Guam, and the U.S. Virgin Islands have separate statutory limits, including a 2026 one-unit ceiling of $1,873,675.
The latest National Association of REALTORS® Existing-Home Sales report was released September 10, 2026 for August sales. Existing-home sales fell 2.0% month over month to a seasonally adjusted annual rate of 3.98 million, the first sub-4-million reading since June 2025. The median existing-home price was $429,100, up 1.6% from a year earlier. The latest home-price figure is not used to determine the $900,000 loan scenario in this article, which is deliberately held constant to isolate jumbo-versus-conforming pricing.
The Federal Reserve’s September 16, 2026 FOMC statement was also reviewed. The Committee voted 12–0 to raise the federal funds target range by 25 basis points to 3.75%–4.00%, its first increase since July 2023, following a divided 9–3 hold at the July meeting. The Committee’s post-meeting projections showed most members expecting one additional quarter-point increase before the end of 2026. The statement cited resilient economic activity and elevated uncertainty tied in part to geopolitical developments, alongside inflation that remains above the Committee’s 2% goal.
All monthly payment and total interest figures are modeled, not measured. We used standard fixed-rate amortization on the stated principal at the stated contract rate, excluding taxes, insurance, mortgage insurance, and HOA costs. Points costs are calculated as the survey-reported points multiplied by loan principal. The second row of the $900,000 comparison table is explicitly hypothetical — conforming pricing is not available above the applicable county loan limit, and the row exists only to isolate the rate-and-points effect from the loan-size effect.
One limitation deserves emphasis. Jumbo qualification standards — credit score minimums, down payment floors, reserve requirements, and debt-to-income overlays — have no single governing agency standard. Every jumbo figure in this article is therefore presented as a market characteristic rather than a universal threshold. Individual lenders can deviate substantially in both directions. Research was last conducted September 2026. Where the latest MBA rate dashboard was newer than the latest fully itemized rate-and-points release, the article labels that distinction rather than inventing a points figure.