When Sabastian Sawe ran 1:59:30 at the London Marathon on April 26, 2026, the headline that wrote itself was "first sub-2-hour marathon." Almost immediately, the obvious objection came back from anyone who had been following the sport for more than a few years: hadn't Eliud Kipchoge already done that in 2019? On the surface, yes — Kipchoge ran 1:59:40.2 in Vienna at the INEOS 1:59 Challenge, on a road, on his feet, with a watch running. On paper, no — that performance was never going to be ratified, and Kipchoge himself never claimed otherwise.
The gap between those two answers is the most-asked question about Sawe's record, and it deserves a real explanation. The short version: Kipchoge's 1:59:40 was a beautifully engineered demonstration that the human body could cover 42.195 kilometres in under two hours. Sawe's 1:59:30 was the first time a human did it under the rules everybody else had to follow. This piece walks through both attempts — and through the World Athletics regulations that separate them.
For the splits and the race-day story of Sawe's run itself, see our full breakdown of the London 2026 record.
The setup: what we mean by "world record"
A world record in athletics is a precise, regulated thing. World Athletics — the sport's global governing body — maintains a set of rules in its Competition and Technical Rules that define which performances are eligible for ratification. The marathon falls under road racing rules, which require, among other things:
- A certified course, measured by a calibrated method (the Jones counter and a standard-bicycle calibration ride) and to a precise length.
- An open competition: any qualifying athlete can enter and contest the race.
- Anti-doping testing in the period before, during and after the event.
- Pacing rules: pacemakers may run with the field but must start with the field; in-and-out pacing (joining mid-race) is not allowed.
- No pace cars or vehicles ahead of the field that could provide drafting, course information or any other aid.
- Hydration rules: athletes take fluids at official aid stations or from designated personal-fluid tables. They cannot be handed bottles by a moving vehicle.
- Maximum permitted shoe stack height of 40mm and a single rigid plate (in effect since 2020).
Most of these rules sound like fine print until you start engineering an attempt to beat them — and then they become the entire game.
Breaking2: Monza, May 2017
The first organised attempt to put a human under two hours for the marathon was Nike's Breaking2 project. After two years of physiological testing, equipment design and athlete selection, Nike staged a controlled time trial on the Formula 1 circuit at Monza on May 6, 2017. Three athletes were entered: Eliud Kipchoge, Lelisa Desisa, and Zersenay Tadese.
The conditions were as optimised as they could be made. The course was the Monza circuit, a 2.4km loop with one tight hairpin, repeated 17 times. There were rotating teams of pacemakers running in an open-ended V formation in front of and beside the lead athlete, drawn from a pool of more than thirty professional runners. A pace car drove ahead of the formation projecting a line of green lasers onto the road to mark the target pace. Hydration was delivered to Kipchoge by a member of the support team riding alongside on a bicycle.
Kipchoge ran 2:00:25 — twenty-five seconds outside the barrier, more than two and a half minutes inside the world record at the time, and entirely outside the rulebook. Nike never asked for ratification; they didn't need to ask, because every adult in the room knew the answer.
The Breaking2 attempt did two things. It demonstrated that 2:00 was, physiologically, in range. And it set the conceptual template for a private, engineered attempt at the barrier — which is the template Kipchoge and the INEOS team would refine two years later.
INEOS 1:59: Vienna, October 2019
On October 12, 2019, Kipchoge ran the INEOS 1:59 Challenge in Vienna's Prater park. The course was a 9.6km loop on the Hauptallee, the long tree-lined avenue that runs through the park. Conditions were colder, the loop was straighter, and almost everything about the engineering had been improved on Monza. Kipchoge ran 1:59:40.2 — the first time a human had ever covered the marathon distance in under two hours, on his feet, in any setting.
It is one of the most remarkable feats of human endurance ever recorded. It is also not a world record. Here is what was different about INEOS, in detail:
- Forty-one rotating pacemakers ran with Kipchoge, in groups of seven at a time. Five pacemakers in front, two behind, rotating in and out at scheduled intervals. The geometry was specifically designed to break the wind for the lead runner. Drafting effects were, by the most rigorous estimates, worth somewhere between two and six seconds per kilometre at that pace — i.e. a difference that easily covers the ten seconds between Kipchoge's run and Sawe's.
- A pace car drove ahead of the formation projecting a green laser onto the road, exactly as in Monza. World Athletics rules prohibit pacing or course information being provided by motorised vehicles in a record-eligible race.
- Hydration was handed to Kipchoge mid-race by a member of the support team on a bicycle. In a record-eligible race, athletes have to take fluids from official aid stations.
- It was not an open competition. The race was designed for one athlete. There were no other competitors, no rivals, no race against the clock that anyone else could enter. World Athletics rules require a record-eligible performance to be set in open competition.
- It was set up specifically to optimise conditions. The date was chosen for weather, the course for flatness and shelter, the time of day for cool air. None of those things, individually, are illegal. The combination signalled what the event was: an exhibition designed to break a number.
INEOS got what it set out to get. Kipchoge said, in the immediate aftermath, that the goal was to show what was possible, not to set a record. He was right on both counts. World Athletics never moved to ratify the time. Guinness World Records did recognise it under two of their categories: fastest marathon distance (male) and first marathon distance run under two hours. Those two are real records. They are not the marathon world record.
What changed in London 2026
Sawe's 1:59:30 in London was set under a completely different rule regime. Specifically:
- Open competition. London 2026 was a TCS London Marathon — a World Marathon Major with more than 50,000 finishers, an elite men's field of more than thirty athletes, and an elite women's field of similar depth. Anyone with a qualifying time could be in the field. Sawe didn't run alone; he raced.
- Standard pacemaker rules. The lead pacers in London started with the field and ran from the gun. They were entered as competitors. They left the race when their job was done. There was no rotating in-and-out, no second wave of pacers brought in at halfway, no formation engineered for drafting.
- No pace car, no laser, no bicycle. Hydration came from the official elite drinks tables at the standard intervals. The lead vehicles in front of the field were the standard race vehicles — TV bike, race officials — at the standard distance.
- A certified, measured course. London is one of the most heavily-measured race courses in the world. The point-to-point format means it is also subject to the start-finish elevation drop limit and the start-to-finish separation distance limit, both of which World Athletics applies to record-eligible courses.
- Anti-doping testing in the standard race protocol, including post-race blood and urine.
- Legal shoes. Sawe's Adidas Adios Pro Evo 3s came in well under the 40mm stack height limit and used a single rigid plate, complying with the World Athletics shoe regulations updated on January 1, 2026.
Three men finished inside Kelvin Kiptum's 2023 record of 2:00:35. Two went under two hours. None of them needed an asterisk.
"But isn't a 1:59:30 in a real race obviously faster than a 1:59:40 with all that help?"
This is the comparison every runner makes, and the honest answer is: we don't actually know, and we don't need to know. The rules don't exist to tell us how fast someone is on a given day. They exist to tell us which performances are comparable to each other.
A useful analogy: in cycling, the Tour de France time trial winner is not the fastest person who has ever ridden a bike against a clock. The track Hour Record holder isn't either. They are the fastest people who have done it under a specific, comparable set of conditions, and that's what makes the records meaningful. The same logic applies in the marathon. A privately staged, single-athlete time trial with rotating drafters and hydration-by-bike is a different event from a London Marathon. They aren't directly comparable, and the rulebook reflects that.
The estimates that have been published — drafting savings, hydration savings, psychological effects of running alone-against-time vs. in a competitive field — vary widely. They suggest that the engineering benefit of an INEOS-style setup is somewhere between forty seconds and two minutes over a regular race. By any of those estimates, Kipchoge's underlying fitness in 2019 was extraordinary, and Sawe's London time, in a normal race, is in the same physiological neighbourhood. We just can't put them on the same line of the world record list.
Why Kipchoge's INEOS run still mattered
It is also worth saying, plainly, that none of this diminishes what Kipchoge did. Three things from the INEOS attempt are now permanent fixtures in the sport's history:
- The proof of concept. Before October 12, 2019, "human under two hours" was a thought experiment with mathematical models attached. After it, it was an empirical fact. Every subsequent attempt — Kiptum's 2:00:35, Sawe's 1:59:30 — was made by athletes who knew, because they had watched it happen, that the barrier was real but breakable.
- The training and equipment template. The INEOS preparation built the playbook for sub-2-hour racing: long runs at marathon pace, full-distance time trials at race intensity, the carbon-plated PEBA-foam shoe, the chemistry of in-race hydration. Almost all of it has been adopted by the field.
- The psychological frame. Kipchoge's repeated public framing — no human is limited — became part of how the next generation of marathoners talk about training. Whether or not you find that kind of mantra useful, it shifted the conversation away from "is sub-2 possible" toward "how soon."
Kipchoge himself, throughout the build-up to INEOS and in every interview since, has been careful about the distinction. He said the goal was to demonstrate possibility, and he said someone else would eventually do it in a real race. Both turned out to be true.
A short summary
| Attempt | Date | Time | Setting | Status |
|---|---|---|---|---|
| Breaking2 (Monza) | May 6, 2017 | 2:00:25 | Closed F1 circuit, rotating pacers, pace car, hydration-by-bike | Not record-eligible |
| INEOS 1:59 (Vienna) | Oct 12, 2019 | 1:59:40.2 | Closed park loop, rotating pacers, pace car, hydration-by-bike | Not record-eligible (Guinness recognised) |
| Chicago Marathon | Oct 8, 2023 | 2:00:35 | Open WMM race, standard rules | World record (Kelvin Kiptum) |
| London Marathon | Apr 26, 2026 | 1:59:30 | Open WMM race, standard rules | World record (Sabastian Sawe) |
The history of the sub-2-hour marathon is a history of two parallel projects: the engineered demonstrations that proved the time was possible, and the open races that finally produced it. Both lines met on April 26, 2026. The second one is the one in the record book.
Further reading
- How Sabastian Sawe broke the 2-hour marathon barrier — the full splits and pacing breakdown of the London 2026 run.
- Who is Sabastian Sawe? — the Kenyan runner behind the record, his coach, and his training base.
- Inside elite marathon training: what it takes to run sub-2 — the physiological numbers that make a sub-2 attempt possible.
- Super shoes and the sub-2 marathon — the foam-and-carbon technology in every recent record attempt.
- Marathon world records — the full historical progression on the men's and women's sides.