Average marathon time

The average marathon time is about 4:30:49: about 4:19:51 for men and 4:47:18 for women. That is Runsy's unweighted mean of four RunRepeat studies published from 2019 to 2025, the largest covering 107.9M race results. The median finisher runs 4:26:33, and the fastest 10% finish in 3:31:46 or better.

As a pace, a 4:30 finish is 10:18 per mile (4:30 marathon pace). The marathon pace chart turns any time on this page into pace per mile and per kilometre, and how long is a marathon covers what training for one takes.

Converted figures are marked. Everything else is quoted as published.

If you ran 4:45 as a woman, whether that is above or below average depends on which study you ask: slower than the 20-year median, faster than both the 2018 global average and the 2024 US one. Anyone quoting a single average without naming the study behind it is quoting one study and not telling you which.

A contradiction inside one report

The State of Running 2019 gives men's average marathon pace as 6:43 per kilometre, which is 4:43:25 over 42.195 km. Its own text puts men at 4:15:13 in 2001 and about seven minutes slower by 2018, roughly 4:22. We could not reconcile the two, so its sex split is out of the synthesis entirely and only its year-stamped all-finisher figure is in.

Corrected on Oct 10, 2026

An earlier version of this page read that report's 2001 sex split as 2018 figures, and missed that Marathon Statistics 2019 publishes a 2018 split. Men moved from 4:17:54 to 4:19:51 and women from 4:49:49 to 4:47:18. The overall figure did not change.

No single dataset reports finish-time distributions for road running. So instead of picking one number and calling it the average:

1. We collected every published central figure we could trace to a named source with a stated sample size.
2. We dropped anything we could not trace to its origin. A magazine quoting a statistic is not a source; we followed each figure back to the study that produced it.
3. We took the unweighted mean of what survived. Unweighted on purpose: the samples overlap heavily, so weighting by size would hand the answer to whichever study counted the most rows.
4. We printed the range as well as the midpoint. Where the sources disagree, the disagreement is the finding.

The result is Runsy's synthesis of published data, not a measurement of our own.

What the arithmetic cannot fix

Means and medians are being averaged together. The percentile study reports a 50th percentile; the others report an average. Finish times have a long slow tail, so a mean sits later than a median, which widens the range and pulls the midpoint slightly late.

Some figures are converted from a published pace. Where a source reports average pace per kilometre, we multiply by the race distance. That is arithmetic on their number, and each converted figure is labelled.

Most inputs share a publisher. The large studies all come from RunRepeat, so their agreement is one research operation agreeing with itself, not independent confirmation.

The peer-reviewed paper on this page is independent of RunRepeat, but it covers a single city's race.

• Percentiles by age. No source we could trace publishes finish-time percentiles broken down by age, with or without sex. The age tables above are averages, and the lookup is for the whole field.
• A global age-by-sex table. The one above is a single Norwegian race. We did not scale it up to the world: that would be a model wearing a citation.
• Anything from parkrun. parkrun's terms prohibit scraping and data reuse, which also rules out the community scrapes and dataset mirrors built from it.
• Strava's Year in Sport. Widely quoted for a median US marathon time. We read the press release and it contains no finish-time statistic at all, so the number circulating under Strava's name is not in Strava's report.
• Any figure after 2024. Not published as far as we can find.