Guide
Alaska, JetBlue and Southwest: the carriers between the extremes
Alaska Airlines runs 78.2% on time, Southwest Airlines 77.1%, JetBlue Airways 72.4%. Three very different operations that get grouped together for no good reason.
Between the three big networks and the ultra-low-cost carriers sits a group that gets described as "the rest" and has almost nothing in common internally. Over June 2025 to May 2026: Alaska Airlines 78.2% on time, Southwest Airlines 77.1%, JetBlue Airways 72.4%.
Cancellations: 1.30%, 0.89%, 2.50%.
Geography explains most of the spread
The single largest factor separating these three is where each one flies, and it is not a comment on how well any of them is run.
An airline concentrated in the Pacific Northwest and along the west coast operates in airspace that produces far fewer delays than the Northeast corridor. An airline built around New York, Boston and Fort Lauderdale is flying into the most congested airspace in the country on one end and hurricane exposure on the other.
Nationally 21% of delay minutes are charged to airspace and volume constraints. Where a carrier chose to build determines how much of that it absorbs, and no operational excellence prices it away.
This is why the comparison worth making is at a single airport rather than nationally, and why the airline-at-airport pages here exist.
Three different structures
A point-to-point network with no hubs distributes disruption differently from a hub-and-spoke one. There are no banks to fall behind, but there is also no single field where spare aircraft and reserve crews are concentrated. Delay propagates along aircraft rotations across the whole map rather than pooling at one airport.
A network built on one dominant coastal hub inherits that hub's constraints entirely. When the hub has a bad day, the whole airline does.
A regionally concentrated network trades reach for consistency. Fewer destinations, less exposure to the difficult airspace, and figures that look better partly because of the map rather than the management.
What the numbers support
That the geography and structure differences are larger than the operational ones, and that any ranking of these three which does not adjust for where they fly is measuring the map.
That the point-to-point model recovers differently rather than better or worse: quicker on ordinary days, slower after a system-wide event, because there is no hub at which to rebuild.
That cancellation rate deserves the same weight as punctuality here as everywhere else, and that the ordering of these three by cancellations is not the same as the ordering by on-time rate.
Using this
If you are choosing between them on a specific route, the route page's comparison table settles it, with the geography confounder removed because both carriers fly the same city pair.
If you are choosing where to concentrate loyalty, look at each carrier's page for your home airport rather than its national figure. A carrier's number at a field where it holds thirty gates and one where it holds three are different products, and only one of them describes your flying.
Why point-to-point recovers differently
Worth understanding, because it is the structural difference that most affects a passenger and it is invisible in an annual average.
A hub-and-spoke airline concentrates its recovery resources at a small number of fields. Spare aircraft, reserve crews and maintenance sit where the schedule converges. When a hub has a bad morning the airline can rebuild from it deliberately, moving passengers across a deep schedule that all passes through the same place.
A point-to-point airline has no such place. Its aircraft move in long chains across the map, each flight handing its aircraft to the next city rather than returning to a base. On an ordinary day this is an advantage: there are no banks to fall behind and a local problem stays local. After a system-wide event it is a disadvantage, because there is no single field at which to reassemble the operation and the disruption propagates along every chain at once.
Neither model is better. They fail on different days, in different shapes, and an annual on-time figure blends both into a number that describes neither.
What to check before choosing
Look at the carrier's page for the airport you actually fly from, not its national figure. A regionally concentrated airline that looks strong nationally may hold three gates at your field with no spare aircraft within five hundred miles.
Then look at the route page for the specific city pair, where the comparison table puts every carrier on that route side by side with the geography confounder removed. That is the only comparison on this site that is fully like for like, and on these three carriers it routinely reorders the national ranking.
Fleet and route length quietly shape the figures
Two more structural factors worth naming, because both move an on-time rate without saying anything about how well an airline is run.
Average stage length: a carrier flying longer sectors carries more scheduled block time in absolute minutes, which absorbs a late pushback that a short-haul operator has to eat. Any comparison across carriers with different route-length profiles is partly measuring geography again.
Fleet commonality: an airline operating one aircraft type can substitute any aircraft for any other, and any crew for any route. A mixed fleet cannot, and a broken aircraft is harder to replace when the substitute needs a differently qualified crew. This is invisible on a booking page and shows up in the cancellation column on a bad day.
Neither factor is a criticism. Both are reasons to prefer the route-level comparison over the national one.