CERN and Mistral AI signed a statement of intent on October 15 to build a Large Physics Model trained on billions of simulated particle collisions. The two sides disclosed no budget or timeline, so for now it is an early signal.
Perspective Coverage
4 publishers
- Builder
- Builder 39%
- Operator
- Operator 19%
- Investor
- Investor 42%
Reality
- Evidence40
- Adoption
- Not scored
- Hype gap+15
- Incentives55
- Confidence70
Six rare beauty-quark decay rates all land slightly off Standard Model predictions in a CMS analysis
CMS measured six rare beauty-quark decay rates in one analysis and found each slightly different from the Standard Model prediction. A separate LHCb angular result also differs a little, and the tensions are far from conclusive.
Reality
- Evidence40
- Adoption
- Not scored
- Hype gap+8
- Incentives
- Not scored
- Confidence50
New wavepacket algorithm brings a simplified particle collision onto a 104-qubit IBM processor
Caltech and University of Washington physicists simulated a particle collision on a 104-qubit IBM chip using a more efficient way to prepare wavepackets. The test used a simplified one-dimensional theory, so it shows the preparation step that had blocked such runs now fits on current hardware for toy models.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+15
- Incentives
- Not scored
- Confidence55
Josep Martí and Pedro Luque-Escamilla say IRAS 18293-0941, an object in a 1983 catalog, is a small black hole in our galaxy with a jet aimed at Earth. If it holds up, it ends a search for a nearby blazar that has run for nearly 35 years.
Reality
- Evidence64
- Adoption
- Not scored
- Hype gap+30
- Incentives
- Not scored
- Confidence60
MIT-led CMS analysis catches individual quarks leaving wakes in quark-gluon plasma
MIT physicists at CERN's CMS report quarks leaving wakes in quark-gluon plasma, calling it the first direct sign the plasma reacts to a quark as a fluid. The wakes bear on a narrower, long-debated question than whether the plasma is a liquid, though the MIT release does not say how strong the signal is.
Reality
- Evidence35
- Adoption
- Not scored
- Hype gap+30
- Incentives55
- Confidence40
Preprint predicts the Penrose process turns neutrons near Sgr A* into PeV protons
Marina Cermeño's team estimates that neutrons decaying near Sgr A* could produce petaelectronvolt protons using energy from the black hole's spin. The gamma-ray and neutrino signals it predicts are too faint for today's detectors and will need upgraded HAWC and IceCube instruments.
Reality
- Evidence35
- Adoption
- Not scored
- Hype gap+30
- Incentives
- Not scored
- Confidence40
CMS measures the diffusion wakes jets leave in quark-gluon plasma, 20 years after the prediction
CMS physicists have measured, for the first time, the diffusion wakes jets leave behind in quark-gluon plasma, 20 years after they were first predicted. The result adds a second test of the fluid picture of the matter that filled the universe for its first microseconds.
Reality
- Evidence60
- Adoption
- Not scored
- Hype gap+10
- Incentives30
- Confidence55
CMS's biggest Higgs combination matches the Standard Model at 5 percent precision
CMS combined 138 fb-1 of 2016-2018 collision data and measured a Higgs signal strength of 1.01 +/- 0.05, against a Standard Model value of 1. Any new physics that alters the Higgs rate now has room mainly near or below that 5 percent precision.
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap+12
- Incentives
- Not scored
- Confidence60
A few days of oxygen beams at CERN gave air-shower models their first laboratory test
Reading a cosmic-ray shower has always meant trusting a computer model of what happens in air. Last July the LHC made the collision those models start from, under conditions physicists chose.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap+10
- Incentives55
- Confidence60
CERN starts pulling the LHC's original focusing magnets to make room for 11.3-tesla replacements
CERN crews have begun dismantling the LHC around ATLAS and CMS to swap its focusing magnets for niobium-tin ones about 40% stronger. The first replacement quadrupole is not due in the tunnel until the start of 2029.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap+10
- Incentives45
- Confidence60
A classical computer confirmed Duke's 13-ion simulation of string-breaking
String-breaking, the moment a stretched bond between two building blocks of matter snaps and new particles appear, has now been observed on a trapped-ion quantum simulator at Duke. Two other groups saw it on different hardware.
Reality
- Evidence62
- Adoption28
- Hype gap+22
- Incentives70
- Confidence60
An empty LHC search narrows where quantum black holes could still be hiding
A search of LHC collision data for microscopic quantum black holes found none. The exclusion limit that follows is the publishable result. Where that limit sits decides whether theorists have to move.
Reality
- Evidence38
- Adoption
- Not scored
- Hype gap+22
- Incentives62
- Confidence45
Radio images catch a Milky Way black hole pointing its jet at Earth
IRAS 18293-0941 is the first microblazar found in our own galaxy. The authors say the spot where its jet strikes a molecular cloud accelerates particles to petaelectronvolt energies, matching a gamma-ray signal there.
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap+18
- Incentives55
- Confidence58
ATLAS measures proton-oxygen collisions more precisely than the shower models differ
The LHC spent part of July 2025 colliding protons with oxygen nuclei to imitate a cosmic ray hitting air, and ATLAS's measurement is now finer than the disagreements among the seven codes used to model showers.
Reality
- Evidence62
- Adoption38
- Hype gap0
- Incentives45
- Confidence55
ATLAS reports entanglement between the Z bosons produced in Higgs decays
Entanglement had been confirmed in photons, electrons and trapped ions, all relatively long-lived systems. An Oxford-led team has now found it in Z bosons made inside the LHC's most violent collisions.
Reality
- Evidence55
- Adoption
- Not scored
- Hype gap+30
- Incentives70
- Confidence52
UCSB physicists publish a null LHC black hole search as an exclusion limit
Tamas Vami and Danyi Zhang looked through CMS data for microscopic black holes, found none, and published the bound. They also describe the search itself as a new way to look for new particles.
Reality
- Evidence57
- Adoption
- Not scored
- Hype gap+15
- Incentives60
- Confidence52
ATLAS finds strong evidence of entanglement between two Z bosons from a Higgs decay
The two bosons decay before any instrument can touch them, so the team read their spins from the angles of the leptons left behind in 13 trillion electron volt proton collisions. Physical Review Letters has the result.
Reality
- Evidence58
- Adoption
- Not scored
- Hype gap+22
- Incentives66
- Confidence56
A drop in J/psi production at sub-proton scales strains conventional nuclear shadowing
ALICE turned near-miss lead collisions into a photon beam and scanned momentum transfer down to a resolution of 0.2 femtometers, about a quarter of a proton. The drop in J/psi production it found there was not expected.
Reality
- Evidence58
- Adoption
- Not scored
- Hype gap+25
- Incentives62
- Confidence55
The equilibrated core outweighs the corona in oxygen collisions past 20 charged particles
A Sophia University calculation of 5.36 TeV oxygen-oxygen collisions at the LHC counts how much of the matter actually equilibrates, and finds a nonequilibrated remainder near 30% that survives even the most central events.
Reality
- Evidence54
- Adoption
- Not scored
- Hype gap+9
- Incentives61
- Confidence58
The metric that will decide the next collider is a ratio, not a total
An efficiency ratio can double while the electricity bill keeps climbing. That gap is where the case for the next collider gets argued, and won.
Reality
- Evidence56
- Adoption28
- Hype gap+22
- Incentives62
- Confidence50