A 1% bound lives inside one lensing model

GW231123 entered the LIGO–Virgo–KAGRA catalogs as the most massive binary black hole merger reported so far, with a total mass of 190–265 solar masses at 90% confidence. The 25 August 2026 Astrophysical Journal Letters paper by Goyal, Villarrubia-Rojo and Zumalacárregui, ApJL 1008 L12, treats that high observed mass as something cosmological redshift plus gravitational magnification can produce if the source sits behind a lens.[1]

The authors add wave-optics diffraction from a compact microlens sitting in an external gravitational potential. Under that embedded point-mass setup, they say the data favor a lensed reading, with a conservative false-alarm probability of about 1% or less. That number is a bound on Bayes factors inside the model they fitted. The bound is a Bayes-factor ceiling inside that fitted model. No lens galaxy is optically identified.[1]

Masses move if the model holds; distance still slides

If that lensing reading holds, the inferred total source mass shifts to 100–180 solar masses, and the compact microlens sits in a 190–850 solar-mass range. The same fit reconstructs a projected offset and the amplitude and orientation of the external potential. Those are outputs of the waveform-plus-lens model, not a second, independent mass scale.[1]

The reconstruction is complete only up to a mass-sheet degeneracy: distance and projected density can trade. Assuming a single galaxy-scale macrolens as a singular isothermal sphere, the authors infer magnification that would place the source at redshift about 0.7–2 and quote about a 55% probability of an additional detectable image. A quieter alternative is that an unlensed, heavier binary with a different spin and waveform family still describes the same strain. The 1% bound does not close that door; it ranks one parameterized lens against the noise and the unlensed templates they compared.[1]

A second image would be the referee's clock

The paper itself names the next measurement that would bite: continued searches for additional macroimages, and for diffraction by stellar fields. Until a second image appears with a delay and magnification the macrolens model can be scored against, the 1% figure remains a within-model bound on one November 2023 strain. One study is a hypothesis wearing a catalog mass.[1]