Real-time or end-of-day rates data: what does your model need?
Last updated: 9 October 2026
Short answer: most finance teams do not need tick-by-tick data. Valuations and month-end reporting need one end-of-day curve with a stated snap time; models need a recent quote when they run; loan systems need daily fixings plus an end-of-day curve. Overnight fixings such as SONIA, €STR and SOFR publish once a day, the next business day.
On evaluation calls, the question comes in many forms. "Is it live or end-of-day? Every 30 seconds? Could we just poll every 15 minutes?" "When is the next update published, is it T+1?" "Why does one index update daily when the others are live?" "Can I get the Monday end-of-day print and compare it with the previous Monday?" The people asking are treasurers, valuation teams, auditors and developers building products on rate data, and the right answer depends on what the number is used for. This guide sets out the terms, the timings and what each use case needs.
What is the difference between real-time, intraday, end-of-day and T+1 data?
The four terms describe how often a value updates and when you can see it. Real-time means each new quote as it arrives; intraday means snapshots through the day; end-of-day means one value per day at a stated time; T+1 means published the next business day.
| Term | What you get | Typical example |
|---|---|---|
| Real-time (live) | Each new market quote as it arrives, or a refresh every few seconds | A swap rate on a dealing screen |
| Intraday | Snapshots during the day, for example every 30 seconds, every minute or every 15 minutes, usually with history kept | A 5Y OIS rate at 09:00, 12:00 and 16:00 |
| End-of-day (EOD) | One value per day, taken at a stated snap time | The 16:00 London close of a SONIA swap curve |
| T+1 | A value for day T published on the next business day | SONIA, €STR and SOFR overnight fixings |
"Live" on a vendor's website rarely means tick-by-tick. Ask for the refresh interval in seconds and whether every point carries a timestamp. A refresh every 30 seconds is fast enough for any modelling or pricing use outside a trading desk.
Why do overnight fixings update once a day when swap rates are live?
Because a fixing and a swap quote are different things. A fixing is a benchmark calculated after the day's transactions are in, so it is published once, the next business day; a swap or OIS rate is a quote that trades continuously.
An overnight fixing measures what was actually paid on overnight money on day T. The administrator cannot calculate it until day T has finished and the transactions are reported, so it is published the following morning:
| Fixing | Administrator | Published | Source |
|---|---|---|---|
| SONIA | Bank of England | 09:00 London on the next London business day | Bank of England |
| €STR | European Central Bank | 08:00 CET on the next TARGET2 business day | ECB |
| SOFR | Federal Reserve Bank of New York | About 08:00 New York time on the next business day | New York Fed |
So on a Friday morning the latest €STR you can get is Thursday's. That is not a stale feed; it is how the benchmark works. You can see the published series on our €STR fixings, SONIA fixings and SOFR fixings pages, and the compounded versions used for loan interest on pages like compounded €STR.
An overnight index swap (OIS) is a different instrument. It exchanges a fixed rate for the compounded fixing over a term such as 5 years, and dealers quote it all day as expectations for future policy rates change. "OIS market data" or "€STR rate data" can therefore mean two things: the daily fixing history, or the intraday OIS curve built from swap quotes. The curve is what you discount and project with; the fixings are what you accrue past interest with. If you are new to the instrument, start with what overnight index swaps are.
Term IBOR-style fixings such as EURIBOR are also published once a day, but on the same business day they are fixed (11:00 CET for EURIBOR) rather than the next. Either way, a fixing is one number per day.
What is a snap time, and why does it matter?
A snap time is the moment at which an end-of-day value is taken. Two providers with identical market data will show different end-of-day curves if one snaps at 16:00 London, another at 16:00 New York and a third at midnight UTC, because the market moves in between.
There is no single industry close for swap curves. Common conventions are:
- 16:00 or 17:00 London for sterling and often for euro curves, and for many European reporting processes
- The local market close for each currency, for example late afternoon for euro, late afternoon New York for dollar
- 15:00 New York, which some US processes use
- The last quote of the day, whenever that market stops updating
A timestamp matters for reconciliation because it lets you compare like with like. If your bank's valuation says the 5Y rate was 4.70% and your model says 4.79%, the first question is when each number was taken. Snap time is the most common cause of a gap between two sources, ahead of conventions and interpolation, as we explain in why swap rates differ between sources. A curve without a timestamp cannot be reconciled against anything.
"When you get a quote from the bank, ask them for a specific time for the quote."
How much can a swap rate move within one day?
Several basis points on an ordinary day, and more on a data release or central bank day. In the illustrative day below, the 5Y dollar swap rate falls 10bp between the London lunchtime and the New York afternoon, so a snap time choice alone can change a valuation materially.
Worked example: an illustrative day of 5Y swap rates
The table below is a hypothetical Thursday in October, when London is on British Summer Time (UTC+1) and New York on Eastern Daylight Time (UTC-4). It shows 5Y spot-starting OIS rates on round illustrative levels to show how intraday moves, market closes and snap times interact. Illustrative figures, not market data.
| UTC | London | New York | 5Y SONIA | 5Y €STR | 5Y SOFR |
|---|---|---|---|---|---|
| Previous day close | 4.000% | 3.000% | 4.500% | ||
| 08:00 | 09:00 | 04:00 | 4.030% | 3.035% | 4.540% |
| 11:00 | 12:00 | 07:00 | 4.060% | 3.045% | 4.580% |
| 15:00 | 16:00 | 11:00 | 4.050% | 3.040% | 4.560% |
| 16:30 | 17:30 | 12:30 | 4.055% (close) | 3.040% (close) | 4.540% |
| 20:00 | 21:00 | 16:00 | 4.055% (no new quotes) | 3.040% (no new quotes) | 4.480% |
| 23:59 | 00:59 next day | 19:59 | 4.055% (no new quotes) | 3.040% (no new quotes) | 4.490% (close) |
What the table illustrates:
- Much of the move can come early. Sterling and euro are 3bp to 3.5bp above the previous close by 09:00 London, before many reporting processes have started.
- Markets close at different times. Sterling and euro quotes stop updating in the late afternoon London time, so any request after that returns the last quote. Dollar quotes carry on into the evening.
- A snap time choice can open a gap. In this example, 5Y SOFR peaks at 4.58% at 12:00 London and falls to 4.48% by 16:00 New York. A 16:00 London snap gives 4.56%; a 16:00 New York snap gives 4.48%. That 8bp gap is a snap time choice, not a data error.
- What 8bp is worth. A 5-year $100m SOFR swap, priced at the 30 September 2026 close, had a PV01 of about $44,100 per basis point, so an 8bp difference in snap time would move the mark-to-market by about $350,000.
- Fixings lag by a day. On the Friday morning after this Thursday, the latest SONIA, €STR and SOFR fixings available would be Thursday's: T+1 publication in practice.
Which frequency does each use case need?
Match the frequency to how the number is used. Reporting needs a fixed, documented close; modelling needs a recent quote; only products that display rates to clients need anything close to live.
| Use case | Frequency needed | Why |
|---|---|---|
| Month-end and quarter-end valuations, hedge accounting | End-of-day, fixed snap time | Auditors need the same curve to come back every time they ask, with a timestamp and a documented source |
| Swap mark-to-market for lenders or counterparties | End-of-day, matched to the counterparty's snap | Reconciling with a bank valuation is only possible on the same date and time |
| Weekly or monthly comparisons ("Monday close vs last Monday") | End-of-day history | You need the same snap on every date, kept unchanged |
| Treasury forecasting and budget models | Daily, or intraday on demand | Rates rarely move enough within hours to change a budget; refreshing when the model is opened is enough |
| Project finance models before financial close | Intraday on demand | Sizing and indicative hedges need today's curve, not yesterday's |
| Checking a bank's swap quote or closing a hedge | Intraday, refreshed within minutes | Rates can move several basis points in a session, so a stale curve gives a misleading benchmark |
| Loan management and accrual systems | Daily fixings (T+1), plus end-of-day curves | Interest accrues on published fixings; forward projections use a daily curve |
| Fintech products and client dashboards | Intraday, every 30 seconds to 15 minutes | Users expect a current number; a timestamp shown next to it avoids disputes |
| Trading and execution | Real-time, tick-by-tick | Only here does every second count, and this is a dealing-screen job |
Polling every 15 minutes is enough for almost every use case on this list. What matters more is that each value carries its timestamp, and that you can ask for the value at a time you choose.
"If you take yesterday's date, then it's locked. It'll use the end of day data as of that date…"
Does FX spot or bond yield data work the same way?
Not always. Many providers hold FX spot rates and government yields as one value per business day, so "FX at 14:00" may not exist as a stored point even when swap curves have full intraday history.
If your process needs FX at a specific time of day, for example a 16:00 London rate for a fund, or the same time as a swap valuation, ask whether the provider stores intraday FX history or only a daily value, and at what time that daily value is taken. Mixing a 16:00 swap curve with an FX rate from a different time introduces a small mismatch that is hard to explain to an auditor later.
How to check this yourself
- Ask for the refresh interval in seconds, and whether "live" means streaming, polling or a delayed feed.
- Ask what "end of day" means for each currency: a fixed clock time, or the last quote before the market stops updating.
- Check that every data point has a timestamp and a stated time zone.
- Request the same date twice, weeks apart, and confirm the historical value has not changed.
- Request a value at a specific past time (for example 16:00 London on your last quarter-end) and check that the response shows the timestamp of the quote it used.
- For fixings, confirm the publication lag: the latest value on a given morning should be the previous business day's.
- Match your counterparty's snap before comparing a valuation, then look at conventions and interpolation.
Where BlueGamma fits
- Swap and OIS curves refresh every 30 seconds in the web app, the Excel add-in and the interest rate API, and every point is timestamped in UTC at source.
- Swaps and OIS keep intraday history. End of day is the last quote before that market stops updating: late afternoon London time for GBP and EUR (recently about 16:00 to 16:30 UTC), and into the evening for USD; daily history takes the last quote before midnight UTC. You can pass your own valuation time and get the last quote at or before it, so you set your own cut per currency.
- FX spot and government yields hold one value per business day.
- Point-in-time curves for any past valuation date, and history is never rewritten, which suits quarter-end valuations.
- The curve-by-curve methodology ships with every trial and can be cited by auditors. For the wider evaluation, see how to choose an interest rate data provider.
If you want to test snap times against your own valuations, start a free 14-day trial or book a call.
Frequently Asked Questions
Do I need real-time interest rate data for month-end reporting?
No, month-end and quarter-end reporting needs one end-of-day curve per reporting date, taken at a stated snap time and kept unchanged. What auditors check is that the curve is timestamped, comes from a documented source and is applied the same way every period. See our guide to quarter-end valuations with forward curves.
When is €STR published?
The ECB publishes €STR at 08:00 CET on each TARGET2 business day, for the previous business day's activity. So the €STR you see on a Friday morning is Thursday's rate. SONIA follows the same pattern at 09:00 London and SOFR at about 08:00 New York time. The history is on our €STR fixings page.
What is the difference between €STR fixing data and €STR OIS market data?
The fixing is one published rate per day for overnight euro money, while OIS market data is the intraday curve of swap rates that exchange a fixed rate for compounded €STR over terms from one week to decades. Fixings are used to accrue past interest; the OIS curve is used to project and discount future cash flows. Our explainer on overnight index swaps covers the instrument.
Is polling a rates API every 15 minutes enough?
For almost every use outside trading, yes. Models, dashboards and loan systems rarely need fresher data than that. What matters more is that each value comes with the timestamp of the quote it was built from, and that you can request a value at a specific past time for reconciliation.
Why does my bank's swap valuation differ from my own end-of-day curve?
The most common reason is a different snap time. On a volatile day a 5-year swap rate can move 8bp or more between 16:00 London and 16:00 New York, which on a $100m swap is several hundred thousand dollars of mark-to-market. Once snap times match, check conventions and interpolation, as set out in why swap rates differ between sources.
Does BlueGamma provide intraday and end-of-day rates data?
Yes. Swap and OIS curves refresh every 30 seconds in the web app, Excel and the API, with every point timestamped in UTC and intraday history kept. You can pass a valuation time and get the last quote at or before it, or take the end-of-day close for each currency. The methodology explains how each curve is built.
.png){kind=logo}