We still don't have a more precise value for "Big G"

It was Henry Cavendish in 1798 who achieved the first direct laboratory measurement of the gravitational attraction between two bodies using a torsion balance, although his target was the Earth’s density. This consisted of a large dumbbell with two-inch lead spheres on either end of a six-foot wooden rod suspended by a wire at its center so it could rotate. There was also a second dumbbell with two 12-inch lead spheres, each weighing 350 pounds, that would attract the smaller spheres when brought close, causing the suspended rod to twist. Cavendish painstakingly recorded those oscillations to measure the gravitational force of the larger spheres on the smaller ones, and from that he could infer Earth’s density. His torsion balance has since become something of a workhorse for physicists keen on refining the value for Big G. The NIST scientists also added an extra twist: They ran two versions of the experiment, one with copper masses and one with sapphire masses, achieving nearly identical values for both. This ruled out the possibility that the specific materials used were affecting the measurements. After all that, they came up with a value of 6.67387×10-11 meters3/kilogram/second2. That’s 0.0235 percent lower than the original BIPM result. Some might question why physicists continue to try to measure the value of G with more precision. One benefit is that it leads to ever-better instruments for measuring small forces, torques, and other subtle effects, advances that benefit science in general. But also, “Every measurement is important, because the truth matters,” said co-author Stephan Schlamminger, a physicist at NIST. “For me, making an accurate measurement is a way of bringing order to the universe, whether or not the number agrees with the expected value.” Metrology, 2026. DOI: 10.1088/1681-7575/ae570f (About DOIs).