Snapshot, not a live feed
Every figure below was pulled on and is not updated afterwards. For where the market stands right now, see the Radar.
The Treasury publishes a daily yield for eleven maturities, from one month out to thirty years. On 1 June 2023 the 3-month bill was the highest-yielding of all eleven. On 29 September 2026 only the 1-month bill pays less than it does. Nothing about the bill itself changed — it is the same contract, with the same guarantee, from the same issuer. What changed is everything it sits next to. That is why “is now a good time to buy a T-bill” has an answer that does not require predicting anything.
Start here: what a Treasury bill actually is
A Treasury bill is a loan you make to the US federal government for a year or less. It works differently from most things that pay interest, and the difference is worth thirty seconds.
A bill pays no interest. You buy it for less than it is worth and you are paid the full amount at the end. A 13-week bill with a face value of $10,000 might cost you $9,895 today; thirteen weeks later the Treasury sends you $10,000. The $105 difference is the interest. That is what “sold at a discount” means, and it is why a bill has no coupon, no payment schedule and nothing to reinvest along the way.
Bills, notes and bonds are the same promise over different lengths. Bills run 52 weeks or less. Notes run two to ten years and pay interest twice a year. Bonds run twenty or thirty. All three are backed by the same issuer, so they carry the same credit risk — which is the lowest available in dollars. They do not carry the same price risk, and that distinction is the whole article.
Bills come in six lengths — four, eight, 13, 17, 26 and 52 weeks. The short ones are auctioned every week, the 52-week every four weeks. At TreasuryDirect the minimum is $100 and you buy in $100 steps, so this is not a large-account product. A brokerage will also sell you bills on the secondary market, usually in $1,000 steps.
One more thing that matters and never appears in the yield: interest on a Treasury bill is exempt from state and local income tax. It is fully taxable federally. If you live somewhere with a high state income tax, a bill at 4.25% can keep more of itself than a bank CD or a prime money-market fund paying slightly more.
The same curve, three years apart
Put both dates on one picture and the question answers itself. In June 2023 the line slopes down: the less time you committed, the more you were paid. Today it slopes up, the way a yield curve normally does.
This is the single most useful chart for anyone deciding how long to lend money. In 2023 committing for ten years cost you 1.89 points a year in forgone yield, so a bill was both the best-paying Treasury and the safest one to own. Today committing for ten years pays you 1.01 points a year. The bill is still the safest. It is no longer the best paid, and those two facts have come apart.
What the bill itself did
The headline rate on cash is down a long way from its peak — and has turned back up.
The low was 3.62%, touched on 18 and 19 December 2025 and again on 7, 8 and 9 January 2026. A bill buyer today is getting 63 basis points more than anyone who bought nine months ago, and 138 basis points less than anyone who bought at the October 2023 peak. Both are true, and which one feels important depends entirely on which number you anchored to.
The price of staying short, in one line
Subtract the 3-month bill from the 10-year note and you get the number that decides what a bill costs you. When it is negative, a bill pays more than a ten-year loan and you give up nothing to stay liquid. When it is positive, that is the annual rent on your flexibility.
Everything written about T-bills between 2022 and 2025 was written inside the dip in that chart, when cash paid the most and risked the least. That was not normal. It was the most inverted the curve had been in four decades, and it has now undone itself.
The bill didn’t get worse. Everything else got better paid.
The four questions that actually decide it
None of these requires a forecast. All four are answerable from numbers published every afternoon.
1. When do you need the money? This is the one that settles most cases, and it has nothing to do with the curve. If you need the cash inside a year — a tax bill, a house deposit, a buffer you might actually draw on — a bill matures when you need it and hands back exactly the amount printed on it. A 10-year note paying 5.26% does not help you in March, because in March you have to sell it to someone, at whatever price exists that morning. A bill held to maturity has no price. It has a date.
2. What is the curve charging you? The 10-year-minus- 3-month spread is the fee, and it is +1.01 points. Lending for three months instead of ten years costs you about a point a year right now. In 2023 that same choice paid you 1.89 points a year, which is why cash was an easy call then and is a real trade-off now.
3. Which way is your reinvestment going? A bill matures fast, so you will buy another one, and the rate you get is whatever exists then. The curve tells you which direction the market has priced. The fed funds effective rate is 3.88%; the 1-month bill is 4.04%, the 3-month 4.25%, the 6-month 4.36% and the 1-year 4.58%. Short lending now pays more than overnight money and more the longer you go. That shape appears when the market expects the overnight rate to be higher later. In 2024 and early 2025 it was the other way round — bills sat below fed funds, cuts were priced, and each roll paid less than the last. Today the roll is priced to improve, not decay.
4. Does the real yield clear your bar? 4.25% nominal minus 3.35% inflation for the year to August 2026 leaves 0.90 points of actual purchasing power. At the bill’s October 2023 peak the same arithmetic gave 2.38 points, because 5.63% nominal met 3.25% inflation. The headline barely moved from “over 4%” to “over 5%” and back, but what you keep fell by more than half.
The curve is evidence about what the market has priced, not about what will happen. The Fed raised its target range to 3.75–4.00% in September 2026 — its first increase since July 2023 — and meets again on 27–28 October. A curve priced for higher rates has been wrong before, in both directions. Bill yields are also pushed around by how much the Treasury happens to be issuing, and comparing a bond-equivalent yield to an overnight rate is not quite apples to apples. These are readings, not predictions.
$10,000 for one year, at today’s prices
The same money, lent for different lengths, at the yields quoted on 29 September 2026. The spread between the shortest and the longest is $101 a year on $10,000 — about a dollar a week.
A hundred dollars a year is what the whole argument is worth on a $10,000 position. On $250,000 it is $2,525 a year, which is why this choice gets more agonising the more money is involved — and why, below some balance, matching the maturity to when you need the cash is simply the better use of your attention.
What this means for your money
A 3-month bill bought on 29 September 2026 returns 4.25% annualised, with certainty, in an instrument that matures in thirteen weeks and owes no state income tax. That is a good absolute deal and it is the worst relative deal in roughly three years. Both halves are the point.
The 1.01 points the 10-year pays over the bill is not free money — it is the fee for taking price risk. If the curve is right that rates are heading higher, that is precisely the circumstance in which a long note loses value, and a buyer who needs the cash early eats the loss. The bill’s weakness and its strength are the same property: it is too short to go wrong and too short to pay much.
The behaviour worth noticing is on the other side. Money-market fund assets reached $8.44 trillion on 1 April 2026, up 12.8% from $7.48 trillion a year earlier — a quarterly figure, five months older than the yields on this page. Money kept arriving in cash while the yield on cash fell 138 basis points from its peak and the rest of the curve overtook it. Whatever the right answer is, the amount of money choosing cash is not evidence about it.
If you came here from our survey of 54,860 box spreads, this is the other side of that coin: a box spread is a synthetic Treasury bill, so the bill’s rate is the ceiling on what the box can pay, and the bill gets there in one trade instead of four. And if you sell cash-secured puts, the collateral sitting behind each put is the same cash a bill would hold — worth checking what your broker pays on it, because at 4.25% the gap is not rounding.
The short version
- On 1 June 2023 the 3-month bill yielded more than all ten other published Treasury maturities. On 29 September 2026 only the 1-month bill yields less.
- The 10-year-minus-3-month spread went from −1.89 points to +1.01, a 2.90-point swing — though still short of the +2.09 seen in May 2022.
- The bill is at 4.25%, down 138 basis points from its 5.63% peak in October 2023 and up 63 from its 3.62% low in December 2025.
- After 3.35% inflation the bill leaves 0.90 points of real yield, against 2.38 points at its nominal peak.
- Bills now yield above the 3.88% fed funds rate and rise with maturity, which is the shape that appears when the market has priced higher rates ahead — so reinvestment currently looks like upside rather than decay.
- Horizon settles most of it: money needed inside a year matches a bill whatever the curve is doing, because a bill held to maturity has a date rather than a price.
Data appendix
Every figure in this post, with the date it was read and the series it came from. Treasury yields are constant-maturity rates and the 29 September 2026 readings were the latest available across all maturities when this was written.
| Metric | Value | Period | Source |
|---|---|---|---|
| 3-month Treasury bill | 4.25% | 29 Sep 2026 | FRED DGS3MO |
| 1-month Treasury bill | 4.04% | 29 Sep 2026 | FRED DGS1MO |
| 6-month Treasury bill | 4.36% | 29 Sep 2026 | FRED DGS6MO |
| 1-year Treasury bill | 4.58% | 29 Sep 2026 | FRED DGS1 |
| 2-year Treasury note | 4.89% | 29 Sep 2026 | FRED DGS2 |
| 10-year Treasury note | 5.26% | 29 Sep 2026 | FRED DGS10 |
| Fed funds effective rate | 3.88% | 29 Sep 2026 | FRED DFF |
| 10-year minus 3-month spread | +1.01 points | 29 Sep 2026 | FRED T10Y3M |
| Same spread, deepest inversion | −1.89 points | 1 Jun 2023 | FRED T10Y3M |
| Same spread, recent high | +2.09 points | 2 May 2022 | FRED T10Y3M |
| 3-month bill, highest daily close | 5.63% | 6 and 12 Oct 2023 | FRED DGS3MO |
| 3-month bill, lowest daily close since | 3.62% | 18–19 Dec 2025 and 7–9 Jan 2026 | FRED DGS3MO |
| 3-month bill rank among 11 maturities | 1st of 11 → 10th of 11 | 1 Jun 2023 → 29 Sep 2026 | FRED DGS1MO–DGS30 |
| CPI inflation, year over year | 3.35% | Aug 2026 | FRED CPIAUCSL |
| Real 3-month bill yield | +0.90 points | 29 Sep 2026 vs Aug 2026 CPI | Derived, FRED DGS3MO and CPIAUCSL |
| Real 3-month bill yield at nominal peak | +2.38 points | Oct 2023 | Derived, FRED DGS3MO and CPIAUCSL |
| Money-market fund assets | $8.44 trillion | 1 Apr 2026 | FRED MMMFFAQ027S |
| Same, a year earlier | $7.48 trillion (+12.8%) | 1 Apr 2025 | FRED MMMFFAQ027S |
| Fed target range | 3.75–4.00% | Raised Sep 2026 | federalreserve.gov |
| Next FOMC meeting | 27–28 Oct 2026 | Scheduled | federalreserve.gov |
| Bill maturities and minimum | 4, 8, 13, 17, 26, 52 weeks; $100 | Current program | TreasuryDirect |
Newer post
Treasury yields fell. Mortgage rates went up anyway.The 10-year dropped 22 basis points. The 30-year mortgage rose 29. The lender's margin, not the government's borrowing cost, is what moved.Earlier post
We priced 54,860 box spreads. The guarantee is real. The profit isn't.Every American-style box we priced locked in a loss — all 21,713 of them. The best credible box anywhere beat a Treasury bill by 0.22 percentage points.Get new posts by email
One email when a new post goes up. No account needed, no other mail, and every message has a one-click unsubscribe.
Educational information only — not investment, tax, or legal advice. Nest Money is not a registered investment adviser or broker-dealer. Market commentary describes what has already happened; it is not a forecast and not a recommendation to buy or sell any security. Options involve substantial risk of loss.