Flight Price Predictor: Will My Flight Price Drop? (2026)

A flight price predictor tells you what fares on your route have done before. It cannot tell you what a sale nobody has filed yet will do.

That gap is the whole story. A forecast is a bet you place before you buy, and if it goes the wrong way, the money is simply gone.

A fare that falls after you book is a different situation entirely, because on most standard US tickets the difference is still claimable.

What is a flight price predictor, and what does it claim?

A flight price predictor reads the price history of a route and turns it into one of two words: buy, or wait.

The four tools most people mean by the term phrase that advice differently, and the differences matter more than the branding does.

• Google Flights price insights label the current fare low, typical or high against past trends. Its help pages say the tip "Prices are unlikely to drop before you book" appears when Google predicts with a high degree of confidence that the fare will not fall before departure.
• Hopper returns a buy-or-wait call with a predicted price, and keeps watching the trip afterward. Its research site says the average user starts watching a trip three to four months out.
• KAYAK Price Forecast predicts whether the fare will rise or fall over the next seven days, and shows the statistical confidence behind the call.
• AirHint is the independent one, built for this question and nothing else. It publishes a probability of a future drop and trains separate models per airline.

Google also runs something different in kind: a price guarantee badge on selected itineraries, which is not advice but a financial commitment.

How accurate are flight price predictors?

Here is the honest version. Two of the four publish an accuracy number, two publish none, and none publish the method behind the figure.

Hopper's research site states that it predicts future flight prices with 95% accuracy up to one year in advance of departure. That page is dated March 13, 2019.

AirHint publishes "80%+ prediction accuracy" and says plainly that the accuracy depends on each airline and each flight.

Google and KAYAK publish no percentage. What they publish instead is a caveat, and the caveats are the most useful thing either company says on the subject.

Set the four side by side and the shape is clear enough.

What is missing from that column of percentages is a published methodology, and anyone outside the company checking it.

The one openly published benchmark in this field is old. In 2003, a University of Washington and USC team tested a fare-prediction system called Hamlet against an algorithm with perfect knowledge of future prices.

Across the full simulation, the paper reports the system saved 4.4% of the ticket price averaged over all 4,488 simulated passengers, and 23.8% for the 341 who could have saved anything at all.

That is a 2003 result on 2003 data, not a measurement of any product selling today. It is quoted here because it is the only accuracy figure in this space with its workings attached.

What can a flight price predictor not know?

Start with what a forecast is built from. Google says its prediction rests on "an analysis of price trends of past flights," and every tool here works the same way.

A sale that has not been filed yet is not in that history. Neither is the competitor's answer to it, nor the schedule change that reshuffles a cabin next month.

None of that is a shortcoming in anyone's model. It is simply what forecasting is.

Horizons matter too. KAYAK is specific about its own window, seven days, so anything past that sits outside what the forecast is claiming.

The conditions under which a drop becomes likely are a separate question with better answers, and our guide to the signs that flight prices are about to drop covers them.

Why a prediction is a bet and a price drop is not

Now the asymmetry that should decide what you actually do.

If you wait on a forecast and the fare climbs, nothing gives you yesterday's price back. That decision is spent, and no policy exists to unspend it.

If you book and the fare falls, US carriers generally leave a path to the difference open on standard fares.

Two rules do that work. The first, 14 CFR 259.5, requires airlines to let you cancel without penalty within 24 hours when you book at least seven days before departure.

After that, the major US carriers have largely eliminated change fees on standard fares. Delta, for one, documents no change fees on its Classic and Extra experiences for tickets originating in the US and Canada.

So rebooking the same trip at a lower price returns the gap rather than costing you one. Our repricing guide has the airline-by-airline detail.

Now look again at how Google structured the one product where it puts money behind a prediction.

Google's price guarantee terms say prices are measured between the moment you select the badged option and the departure of your first flight, with the difference paid to your Google Pay account.

Read the shape of that. The company with as much fare data as anyone does not sell certainty before purchase; it picks itineraries it is confident about, then pays what it owes afterward.

The commitment lives after the booking. That is the tell.

Autopilot applies the same logic to your own reservation. On cash tickets booked directly with Alaska, American, Delta, JetBlue, or United, add your trip and Autopilot tracks the fare and reprices the reservation when it drops, with the savings coming back as airline credit or a refund.

Should you book now or wait?

The question has a defensible answer, and it is not that forecasts are worthless.

It is that you should stop asking a forecast to carry a decision it cannot carry, and spend your attention on the side of the booking where outcomes are still reversible.

1. 1

   See where today's fare sits

   Google Flights will tell you whether the price is low, typical or high against the route's own history. Treat that as calibration, not as a promise.
2. 2

   Decide what you would be content to have paid

   If the fare is one you can live with and the seats or times matter to you, waiting for a forecast to resolve buys you very little.
3. 3

   Book directly with the airline where you can

   A ticket bought on the airline's own site is the airline's to service, which is what keeps the 24-hour window and the carrier's own rebooking rules available to you.
4. 4

   Keep the fare tracked after checkout

   The drop you can act on is the one that happens once you already hold the ticket. Our guide to the best flight price tracker covers the tooling side.

For the narrower question of when in the calendar to buy, how far in advance to book flights has the windows by trip type.

One exception changes the arithmetic completely.

Where flight price predictions still earn their keep

None of this makes the forecasts noise, and the conclusion is emphatically not to stop paying attention to fares.

It is that prediction is weak in exactly one place, before you commit, and that is the one place most people ask it to be strong.

• Comparing dates and months. Route history is genuinely informative about seasonality, which is the part of fare behavior that repeats.
• Sanity-checking an unfamiliar route. A low or high label tells you whether a number you have no instinct for is unusual.
• Deciding how hard to shop. A high label with a long runway is a reason to keep looking; a low label is a reason to stop.
• Understanding the direction of travel. Whether fares soften as departure nears is a real question, and the answer runs against most people's instincts.

Used that way, a predictor is a good instrument. Used as an oracle before an irreversible purchase, it is being asked for something none of its makers claim it provides.

Fares will keep moving for as long as airlines keep repricing seats, and no model gets to see the next sale before it is filed.

So put the effort where the outcome can still change. Choose a fare you are content with, then make sure someone is claiming the difference when it falls.