Does Booking Earlier Make Domestic Flights Cheaper? What 481 FlightGPT Searches Show (Jul–Nov 2026)
By Arjun Kapoor (Arjun Kapoor tracks error fares, mileage runs and award-chart sweet spots for Indian travellers. He moderates two Telegram fare-alert channels and has booked Europe round-trips at sub-₹25,000 four times in the last 24 months.) · Published · 11 min read
Across 481 FlightGPT searches on 27 domestic routes between 04 Jul, 2026 and 15 Nov, 2026, the typical fare barely changed with how far ahead people searched. What changed was the spread: searches 0 to 2 days and 60 or more days out had the most expensive results, and searches 14 to 29 days out had the most cheap ones.
The answer in one table
The standard advice is to book early. A Ministry of Civil Aviation release of 21 Jul, 2022 puts it plainly: because of dynamic pricing, tickets bought in advance are much cheaper than those bought near the date of travel. We wanted to see how that looks in the searches people actually ran on FlightGPT. The short version: the middle of the market barely moved, but the odds of a very cheap or very expensive result did.
| Days between search and departure | Searches | Median ratio to route median | Mean ratio | At least 10% below route median | At least 20% above route median |
|---|---|---|---|---|---|
| 0 to 2 days | 160 | 1.00 | 1.118 | 18% | 23% |
| 3 to 6 days | 42 | 1.00 | 1.059 | 12% | 17% |
| 7 to 13 days | 70 | 1.01 | 1.059 | 7% | 13% |
| 14 to 29 days | 110 | 0.954 | 1.148 | 25% | 7% |
| 30 to 59 days | 66 | 1.00 | 1.097 | 11% | 18% |
| 60 days or more | 33 | 1.00 | 1.091 | 21% | 24% |
A ratio of 1.00 means the search found that route's typical fare. A ratio of 0.954 means 4.6% below it. These are past FlightGPT search results, not live prices and not a forecast of what any future search will show.
How we measured it
Why ratios and not rupees
Routes have very different price levels. Averaging raw fares would mostly tell you which routes people searched, not when they searched. Dividing by each route's own median removes that. It also means this study prints no rupee fares at all; for route-level fares, see pages such as Delhi to Mumbai or Delhi to Goa, which carry their own dated search data.
Why we report shares, not just medians
A median tells you what a typical search found. It does not tell you how often a traveller got burned or got lucky. So for each lead-time bucket we also counted the share of searches whose lowest fare was at least 10% below the route median (a good result) and at least 20% above it (a bad one).
What the medians say: the typical fare hardly moved
Five of the six lead-time buckets had a median ratio between 1.00 and 1.01. Searching 0 to 2 days before departure found, at the median, the same fare as searching 60 or more days ahead. Only the 14 to 29 day bucket stood out, at 0.954.
That does not contradict the ministry's statement so much as narrow it. Airline pricing runs in fare buckets, or Reservation Booking Designators, as the same 2022 release explains. In our data that mechanism shows up less as a steady climb towards departure and more as occasional sharp jumps, and a median does not catch occasional jumps. What the median hides is the tail.
The mean tells a different story
Every bucket's mean ratio sat above 1.00, from 1.059 to 1.148. When the mean is above the median, a minority of searches found fares well above typical and pulled the average up. The 14 to 29 day bucket had the highest mean, 1.148, even though only 7% of its searches were 20% or more above the route median. One possible explanation, which we have not confirmed, is a small number of very high results inside that bucket; our summary data does not show the individual searches, so we will not guess which routes they were. The practical point: an average fare figure for a booking window can mislead you. The shares in the next section are more useful.
Where lead time does show up: the odds of a good or bad result
Last minute, 0 to 2 days
This was one of the two riskiest windows. 23% of 160 searches found a lowest fare at least 20% above the route's median, the second-highest share in the study. But 18% found one at least 10% below. Last-minute fares in our data were not uniformly expensive; they were unpredictable in both directions.
Two to four weeks, 14 to 29 days
This was the best window on both measures. 25% of 110 searches came in at least 10% below the route median, the highest share of any bucket, and only 7% came in 20% or more above, the lowest share. If you want to cut the chance of overpaying, this is where our searches point.
One to two weeks, 7 to 13 days
Steady but rarely cheap: 7% of 70 searches were good results and 13% were bad ones. If your trip is ten days away, you will most likely pay about the usual fare.
Far ahead, 60 days or more
The widest spread: 21% good and 24% bad across 33 searches, the highest bad share in the study. Booking two months out did not protect travellers in our sample. With only 33 searches, treat this bucket with the most caution.
What this means for a real booking
Say your usual route has a typical fare you know well. Searching two to four weeks ahead, our data suggests roughly one search in four beats that typical fare by 10% or more, and roughly one in fourteen costs 20% or more extra. Searching in the last two days, nearly one in four costs 20% or more extra. FlightGPT searches one date at a time and has no fare alerts or fare calendar, so the useful habit is to search your date, then ask for the day before or after straight away in the same chat and compare.
Does the day you fly matter more?
We ran the same ratio by weekday of departure.
| Travel day | Searches | Median ratio to route median |
|---|---|---|
| Monday | 71 | 0.952 |
| Tuesday | 59 | 1.003 |
| Wednesday | 41 | 0.997 |
| Thursday | 52 | 1.068 |
| Friday | 76 | 1.016 |
| Saturday | 79 | 1.00 |
| Sunday | 103 | 1.00 |
Monday departures had the lowest median ratio (0.952) and Thursday the highest (1.068), a gap of 11.6 percentage points. That is wider than the gap between any two lead-time medians, which was 5.6 points (1.01 against 0.954). In this sample, which day you flew moved the typical fare more than how early you searched. The samples are small, from 41 searches for Wednesday to 103 for Sunday, so read this as a direction to test, not a rule.
Limitations, and how to use this
What this study cannot tell you
- Small sample. 481 searches over 27 routes is enough to see patterns, not to prove them. The 3 to 6 day bucket has 42 searches and the 60-plus bucket 33.
- Repeat searches. The same visitor may search the same trip several times. Those searches are not independent, and they may cluster in one bucket.
- One source of fares. Every figure is the lowest fare FlightGPT's search returned at that moment, from one fare source. Other channels may have shown different prices.
- Not live, not a forecast. These are past search results. Airlines reprice constantly, and a future search can come back higher or lower.
- No season control. The window from 04 Jul, 2026 to 15 Nov, 2026 includes monsoon months and the festival season. We did not separate them.
- A third of searches were last minute. 160 of 481 searches, about 33%, were 0 to 2 days out, so that bucket carries the most weight.
The regulator's view, for context
Fares in India are set by airlines, not the government. A Lok Sabha reply from the Ministry of Civil Aviation (Unstarred Question 2007, July 2025) says fares follow supply and demand, driven by seat occupancy, fuel costs, aircraft capacity and competition, and that the DGCA's Tariff Monitoring Unit checks fares on 78 randomly selected routes each month, covering about 27% of domestic traffic. A PIB release of 28 Mar, 2022 adds that Air Transport Circular 02 of 2010 requires airlines to publish route-wise tariff sheets on their websites; IndiGo's tariff sheet, for example, is a PDF linked from its website footer. The DGCA's data and reports section publishes domestic air transport data if you want the demand side of the picture.
Practical takeaways
- Aim to search two to four weeks before you fly, where our searches showed the most cheap results and the fewest expensive ones.
- If you must fly within two days, search more than one date and more than one time; the spread is wide in both directions.
- Booking far ahead raises the chance your plans change. Air India's Smart Fares page shows Flex fares rescheduling at no cost and Value fares paying the standard fee; that difference can outweigh a small fare saving.
- Try a Monday or midweek departure if your dates can move.
Sources checked 15 Nov, 2026
- FlightGPT search data: 481 searches on 27 domestic routes with 5 or more searches, 04 Jul, 2026 to 15 Nov, 2026; lowest one-adult economy fare per search divided by route median (first-party, not live)
- PIB, Ministry of Civil Aviation, 21 Jul, 2022: airfare is market driven (Rule 135 of the Aircraft Rules, 1937; fare buckets; advance purchase statement)
- PIB, Ministry of Civil Aviation, 28 Mar, 2022: DGCA monitors airfares (Air Transport Circular 02 of 2010; Tariff Monitoring Unit)
- Ministry of Civil Aviation, Lok Sabha Unstarred Question 2007, July 2025 (pricing factors; 78 monitored routes; about 27% of domestic traffic)
- DGCA digigov portal: data and reports (domestic air transport data; context only, no figures quoted)
- IndiGo: tariff sheet (PDF) (opened to confirm it is published; no fares quoted from it)
- Air India: Smart Fares (change and cancellation terms by fare brand)
Frequently asked questions
Is it cheaper to book domestic flights in India early?
Not at the median in our data. Across 481 FlightGPT searches from 04 Jul, 2026 to 15 Nov, 2026, five of six lead-time buckets had a median ratio of 1.00 to 1.01 against the route's typical fare. Early searching changed the odds of extremes instead.
What is the best time to book a domestic flight in India?
In our searches, 14 to 29 days before departure. 25% of those 110 searches found a fare at least 10% below the route median and only 7% were 20% or more above it. These are past search results, not a forecast.
Are last-minute flights always more expensive?
No, but they are riskier. Of 160 FlightGPT searches 0 to 2 days before departure, 23% were at least 20% above the route median, while 18% were at least 10% below it, between 04 Jul, 2026 and 15 Nov, 2026.
Does booking two months ahead guarantee a low fare?
No. Our 33 searches made 60 or more days ahead had the highest share of expensive results, 24% at least 20% above the route median, alongside 21% at least 10% below. The sample is small.
Which day of the week was cheapest to fly in this study?
Monday. Monday departures had a median ratio of 0.952 across 71 searches, and Thursday the highest at 1.068 across 52 searches. Day of travel moved the typical fare more than lead time did.
How was the FlightGPT lead-time study calculated?
Each search's lowest one-adult economy fare was divided by the median for its route, across 27 domestic routes with 5 or more searches between 04 Jul, 2026 and 15 Nov, 2026. Results were grouped by days to departure and weekday.
Can these numbers predict what I will pay?
No. They are past FlightGPT search results from one fare source, include repeat searches by the same visitors, and do not control for festivals. Airlines reprice constantly; the fare you see when you search is the one that counts.
Does the government cap domestic airfares?
No. A Ministry of Civil Aviation release of 21 Jul, 2022 says fares are market driven under Rule 135 of the Aircraft Rules, 1937, and a July 2025 Lok Sabha reply says the DGCA monitors 78 routes monthly, about 27% of domestic traffic.
Does FlightGPT have fare alerts or a fare calendar?
No. FlightGPT searches 1 date at a time. To compare, search your date and then ask for another date straight after in the same chat, then book the one you want with FlightGPT there.