Guide
The most and least reliable flight numbers in the US
Of 3,252 US flight numbers with a full record, the best arrived on time 98.4% of the time and the worst 36.9%. Here is the full ranking, and why a flight number can beat its own route.

Most delay statistics stop at the route or the airline. That is useful, but it is not what you book.
You book a flight number: one scheduled departure, at one time of day, flown by one operating carrier. Two flights on the same route can have very different records, and the flight number is where that difference shows.
This site tracks 3,252 US flight numbers that flew their main route at least 300 times from September 2025 to August 2026. Across all of them, the median flight arrived on time 81.6% of the time, using the DOT's definition of on time: less than 15 minutes after the scheduled gate arrival.
The spread around that median is wide. The most punctual number in the data, OO 4030, flown by SkyWest Airlines from Palm Springs (PSP) to Salt Lake City (SLC), arrived on time 98.4% of the time. The least punctual, F9 3288 from Atlanta (ATL) to New York (LGA) with Frontier Airlines, managed 36.9%. That is a gap of 61.5 points between two flights sold on the same national network.
The 15 most punctual flight numbers
| Flight | Airline | Route | Departs | Flights | On time |
|---|---|---|---|---|---|
| OO 4030 | SkyWest Airlines | PSP → SLC | 06:00 | 365 | 98.4% |
| OO 4037 | SkyWest Airlines | SUN → SLC | 07:30 | 363 | 98.1% |
| OO 3745 | SkyWest Airlines | TUS → SLC | 07:10 | 365 | 97.8% |
| OO 3901 | SkyWest Airlines | HLN → SLC | 05:30 | 365 | 97.8% |
| UA 596 | United Airlines | LAS → DEN | 05:30 | 365 | 97.8% |
| DL 1270 | Delta Air Lines | PDX → SLC | 05:45 | 365 | 97.5% |
| DL 2369 | Delta Air Lines | PDX → MSP | 06:00 | 359 | 97.5% |
| DL 1535 | Delta Air Lines | PHX → SLC | 06:00 | 365 | 97.3% |
| OO 4230 | SkyWest Airlines | GJT → SLC | 07:40 | 365 | 97.3% |
| UA 459 | United Airlines | PDX → DEN | 05:00 | 362 | 97.2% |
| UA 487 | United Airlines | IAH → SFO | 05:22 | 319 | 97.2% |
| UA 2285 | United Airlines | SMF → DEN | 08:50 | 341 | 97.1% |
| OO 4061 | SkyWest Airlines | OAK → SLC | 05:55 | 365 | 97.0% |
| DL 1376 | Delta Air Lines | PBI → ATL | 06:00 | 326 | 96.9% |
| YX 3679 | Republic Airways | PIT → IAD | 06:10 | 360 | 96.9% |
Each row links to the flight's own page, with its schedule, its day-of-week record and its delay causes, and to the route it flies.
A few things stand out in lists like this one. First, 14 of the 50 most punctual flight numbers are flown by regional carriers, the airlines that operate shorter flights under contract for the big brands. Those flights are often the first rotation of an aircraft's day, leaving a small airport early in the morning with nothing earlier in the schedule to inherit a delay from.
Second, look at where they land. The most common arrival airport among the top 15 is Salt Lake City (SLC), with 8 of them. A hub with spare runway capacity at the time a flight arrives can absorb it without a wait, and that is the situation in which a short feeder flight posts a near-perfect record.
The 15 least punctual flight numbers
| Flight | Airline | Route | Departs | Flights | On time |
|---|---|---|---|---|---|
| F9 3288 | Frontier Airlines | ATL → LGA | 20:57 | 363 | 36.9% |
| OO 5961 | SkyWest Airlines | SMF → SFO | 20:30 | 332 | 43.2% |
| UA 254 | United Airlines | EWR → ORD | 18:20 | 306 | 43.5% |
| UA 2262 | United Airlines | RDM → SFO | 13:55 | 363 | 43.9% |
| DL 2267 | Delta Air Lines | LAX → SFO | 13:40 | 364 | 44.3% |
| AA 2186 | American Airlines | DFW → MCO | 18:45 | 350 | 45.3% |
| DL 2879 | Delta Air Lines | ATL → ORD | 17:42 | 344 | 45.6% |
| B6 1354 | JetBlue Airways | DCA → BOS | 16:59 | 320 | 45.7% |
| B6 2754 | JetBlue Airways | DCA → BOS | 17:59 | 335 | 46.1% |
| UA 2497 | United Airlines | IAD → ORD | 17:15 | 365 | 46.5% |
| AA 1322 | American Airlines | DFW → DEN | 16:13 | 324 | 47.4% |
| UA 661 | United Airlines | EWR → ORD | 16:45 | 346 | 47.6% |
| YX 5763 | Republic Airways | DCA → LGA | 17:40 | 347 | 47.8% |
| F9 4401 | Frontier Airlines | LAS → SFO | 14:48 | 328 | 48.0% |
| B6 1798 | JetBlue Airways | FLL → ATL | 20:00 | 304 | 48.1% |
The bottom of the ranking usually looks different from the top. Evening departures are common, because by the evening an aircraft has flown several legs and every delay earlier in its day has been carried forward. Congested destinations are common too, because a flight that arrives into a queue at a busy hub is late through no fault of its own departure.
There are 151 flight numbers in the data that arrived on time less than 60% of the time, against 575 that managed 90% or better. On a flight in the first group, a traveller should expect to be late on more than one trip in three, and I would not put a connection behind any of them.
Why a flight number can beat its route
A route record averages every departure on the route, all day, on every carrier. A flight number isolates one slot. That is why the flight pages on this site so often show a flight several points ahead of, or behind, its route.
Three things drive the gap:
- Time of day. Early departures inherit less. A route that runs 75% on time overall can contain a 6am flight that runs 90% and an 8pm flight that runs 60%.
- Aircraft rotation. A flight that is the second leg of an aircraft's day depends on the first leg arriving on time somewhere else. The site's delay cause figures call this late-arriving aircraft, and it is usually the largest single cause of delay minutes.
- Operator. The same route can be flown by a mainline carrier and by a regional partner, with different crews, aircraft and spare capacity.
Look at any route page and its "Flights on this route" list to see the spread for yourself. The compare page and the search box both take a flight number such as DL 1234.
How the airlines compare on flight numbers
| Airline | Flight numbers | Median on time | Most punctual number |
|---|---|---|---|
| SkyWest Airlines | 454 | 83.5% | OO 4030, 98.4% |
| Delta Air Lines | 1,286 | 83.1% | DL 1270, 97.5% |
| Alaska Airlines | 173 | 81.9% | AS 750, 96.1% |
| Republic Airways | 115 | 81.9% | YX 3679, 96.9% |
| Envoy Air | 62 | 81.7% | MQ 4139, 95.3% |
| American Airlines | 214 | 81.5% | AA 1903, 94.4% |
| United Airlines | 606 | 81.1% | UA 596, 97.8% |
| PSA Airlines | 44 | 75.8% | OH 5386, 91.1% |
| Frontier Airlines | 95 | 73.3% | F9 686, 92.6% |
| JetBlue Airways | 202 | 73.0% | B6 1129, 95.4% |
Ranked by the median flight number, SkyWest Airlines leads at 83.5% on time and JetBlue Airways trails at 73.0%. The median is used rather than the average so that a handful of very bad or very good numbers does not decide the ranking.
How I pick between two flight numbers
Start with the route page for the trip you are planning, because it shows every carrier and every flight number on the route side by side. Then open the flight numbers you are choosing between.
On a flight page, check three things before trusting the headline rate:
- The schedule section. If the flight has been retimed during the year, the rate blends different departure times, and the most recent slot is the one that matters.
- The quarterly trend. A flight that ran 90% on time in one quarter and 65% in another is less predictable than its annual figure suggests.
- The share an hour or more late. For a connection, this matters more than the on-time rate, because a 20-minute delay rarely breaks a connection and a 70-minute one usually does.
If the flight number you want has a weak record, the same route page will usually list an earlier departure or another carrier with a better one. If there is no alternative, build a longer connection and use the connection risk calculator to see how much time the inbound flight's record actually calls for.
What the ranking does not tell you
A flight number record is a base rate, not a forecast. A 95% on-time flight will still be late on some days, and weather does not respect a good track record.
Flight numbers also change. Carriers move numbers between routes during the year, and this site only reports a number on a route where it flew at least 250 times. Always check that the route on your booking matches the route on the flight page before relying on its rate.
Southwest is the extreme case. It reuses flight numbers across different city pairs on different days, so only six Southwest numbers fly one route often enough to earn a page. The largest airline in the data by flights barely appears in these tables, and I would not read that as a verdict on its punctuality.
Finally, the operating carrier matters: a flight flown by a regional partner is listed under that partner's own code and number. The flight page says so where it applies.
Common questions
Is a flight number the same flight every day? Usually, but not always. Carriers retime flights between seasons and sometimes move a number to a different route. This site reports a flight number only on a route where it flew at least 250 times, and each flight page shows every scheduled departure time that was used on at least one flight in ten.
Why does my airline's flight not appear under its own code? Many short flights are operated by regional carriers for a larger airline. The DOT data records the operating carrier, so those flights appear under the regional carrier's code and number. Search by the route instead, and the route page lists every operating carrier.
Does a high on-time rate mean the flight is rarely cancelled? No. The on-time rate counts only flights that operated. Diverted flights are left out of it too, and a flight that returns to its departure airport counts as a diversion in the file. Each flight page shows the cancellation rate separately, and a flight can score well on punctuality while cancelling more often than its route.
How often is the ranking updated? Every month, when the DOT publishes a new month of data, about two months after it ends. My pipeline downloads that month's file, roughly 250 MB zipped, and re-aggregates the latest twelve months with DuckDB; the build refuses to publish a guide quoting a figure it has not recomputed. The names at the top can change each time.