A single ten-month run added 161 gravitational-wave detections and pushed the total ever recorded to 390 — including two mergers that may be black holes built from earlier black holes

Ten years ago, the record of gravitational waves ever detected held exactly one entry: a single chirp from two black holes that collided more than a billion years ago and shook spacetime hard enough to register on Earth in September 2015. The number now stands at 390.

That is the headline figure of the fifth Gravitational-Wave Transient Catalog, or GWTC-5.0, which the LIGO–Virgo–KAGRA collaboration released on May 26, 2026. The catalog added 161 new detections gathered in a single ten-month stretch between April 2024 and January 2025, and it pushed the running total past a threshold that would have seemed absurd a decade ago. Among the new entries are the best-located gravitational-wave source ever found, the clearest signal ever recorded, and two collisions that may involve black holes that were themselves born from earlier black holes.

What the number 390 actually counts

A gravitational-wave catalog is not a photo album. It is a census. Each entry is a moment when two dense objects, almost always a pair of black holes, spiraled together and merged, radiating a burst of gravitational waves: ripples in spacetime itself that stretch and squeeze everything they pass through. By the time those ripples reach the detectors, they move the mirrors at the ends of the instruments’ kilometers-long arms by less than the width of a proton, which is why extracting a signal at all takes years of noise-hunting. The catalog collects the moments when that extraction succeeds and the physical details wrung from each one: the masses of the objects, their distance, and the way they were spinning.

The 161 new events came from the second half of the network’s fourth observing run, a phase the collaboration calls O4b, running from April 10, 2024 to January 28, 2025. Almost all are pairs of black holes, and the collaboration counts them among the 390 confirmed events it has recorded since 2015. That word, confirmed, carries a specific and statistical meaning here, one worth unpacking once the records are in view.

A graveyard filling up faster than anyone expected

The collaboration keeps a chart it calls the “Masses in the Stellar Graveyard,” a scatter of every black hole and neutron star whose mass has been measured. For decades the plot filled slowly, one object at a time, from telescopes watching X-rays and starlight. Gravitational-wave detectors changed the pace. The network has logged its detections in nine and a half years of listening; the electromagnetic observations beside them represent roughly sixty years of searching. The plot, in the collaboration’s own words, is now bursting at the seams.

Part of that speed is the detectors getting better between runs. The fourth observing run alone accounts for about 75 percent of every gravitational-wave event found since 2015. Each upgrade cycle, when the instruments go quiet for commissioning, buys a jump in sensitivity that shows up as a flood of new signals when observing resumes.

The records inside the catalog

A census this large is mostly valuable in bulk, but a handful of the new entries stand out individually.

The best-localized source is GW240615, detected on June 15, 2024. Because all three instruments, the two US LIGO detectors and Virgo in Italy, were listening at once, the network triangulated the source to an area of just six square degrees of sky. That is small enough for optical telescopes to sweep in search of any accompanying flash. A tight fix on the sky is what lets other observatories point quickly enough to catch light from the same event, the kind of pairing that turned a 2017 neutron-star collision into one of the most-studied events in astronomy. It came from two black holes of about 26 and 30 solar masses colliding more than three billion light-years away, which on their own would produce no light for anyone to chase, but the localization record still matters for the day the network catches something that does. The clearest signal is GW250114, which reached Earth on January 14, 2025 with a signal-to-noise ratio of 76.9, the highest yet, from two nearly identical black holes of 32 and 34 solar masses. Its clarity has already fueled separately published tests of general relativity. The strangest entries are GW241011 and GW241110, detected a month apart in late 2024, and they deserve their own section.

Black holes that may have been black holes before

GW241011 and GW241110 came from mergers roughly 700 million and 2.4 billion light-years away. What sets them apart is spin. The way the black holes were rotating, the orientation and rate, carries a fingerprint of their history, and in these two cases the fingerprint points to something recursive: the objects may be second-generation black holes, meaning each was already the product of an earlier merger before it went looking for a partner again.

That would require a specific kind of neighborhood. In most of the galaxy, two black holes that merge are unlikely to ever encounter a third. But in a dense stellar cluster, where black holes sink to the crowded center and jostle for millions of years, a merged black hole can find a new companion and merge again. The growing catalog has enough events now that the collaboration can begin to treat these repeat-merger black holes as a distinct sub-population with shared traits, rather than as one-off curiosities. A companion paper released with the catalog works through that population in detail.

Where the round numbers get slippery

Three things are worth keeping straight before the round number travels.

First, the collaboration reports all 390 as confirmed events, but “confirmed” here rests on a statistical threshold rather than a yes-or-no switch. The catalog paper admits each new signal when at least one search algorithm puts its probability of being a real astrophysical event at 50 percent or higher, and the event survives validation checks. Most clear that bar with room to spare; a handful clear it by less. The tally is solid, but it is built from a spectrum of confidence, which is what a population study needs and what gets flattened when 390 is quoted as 390 identical certainties.

Second, the second-generation reading of GW241011 and GW241110 is an inference from spin, not a picture of a black hole’s family tree. It is the best explanation for the measurements, and it is still an explanation being tested against alternatives.

Third, the catalog sharpened an independent measure of the Hubble constant, the number that describes how fast the universe is expanding, improving its precision by about 25 percent. That is a real gain, but gravitational waves have not resolved the standing disagreement between different methods of measuring cosmic expansion. They have added a cleaner third voice to an argument that is still open.

Read that way, the qualifiers are what let the catalog work at all. A census is only as good as the honesty of the count behind it, and this one is being counted in the open.

From one chirp to a population

There is more of the fourth observing run still to analyze, with the final portion due for public release in December, so the count will climb again before the year is out. But the shift that GWTC-5.0 marks is less about any single number than about what kind of science this has become. In 2015, detecting a black-hole merger at all was the discovery. In 2026, a single collision is a data point, and the interesting object is the crowd: hundreds of dead stars, their masses and spins and distances laid side by side, dense enough at last to ask what the population as a whole is trying to say.