Shomik Adhicary, Pratyusava Baral, Amanda Baylor, Becca Ewing, Yun-Jing Huang, Rachael Huxford, Prathamesh Joshi, James Kennington, Ryan Magee, Cody Messick, Wanting Niu, Cort Posnansky, Surabhi Sachdev, Shio Sakon, Urja Shah, Divya Singh, Leo Tsukada, Zach Yarbrough, Noah Zhang, Kipp Cannon, Sarah Caudill, Bryce Cousins, Jolien D. E. Creighton, Heather Fong, Richard N. George, Olivia Godwin, Reiko Harada, Soichiro Kuwahara, Alvin K. Y. Li, Duncan Meacher, Soichiro Morisaki, Debnandini Mukherjee, Alexander Pace, Anarya Ray, Stefano Schmidt, Ron Tapia, Koh Ueno, Aaron Viets, Leslie Wade, Madeline Wade, Graham Woan, Chad Hanna

Gravitational-wave observations of merging binary neutron stars and black holes are now routinely made by detectors in the Advanced LIGO-Virgo-KAGRA network. Neutron star binary systems may also produce detectable electromagnetic and particle emission over times scales ranging from seconds to years. Real-time gravitational-wave searches play a central role in enabling time-critical electromagnetic and/or neutrino follow-up observations. During the fourth observing run (O4) of the Advanced LIGO-Virgo-KAGRA network, multiple real-time searches operated continuously to identify candidate gravitational-wave events and publicly disseminate information about these discoveries. Here, the performance and results of the GstLAL real-time analysis are reported. The analysis is designed to identify candidates with low latency, high detection efficiency, and sustained operational uptime over long observing periods. Across O4, it produced initial candidate uploads with a median latency of 15.8 s while maintaining an effective uptime of 98% during the first two parts of the observing run. During the run, the analysis contributed to 250 candidates classified as astrophysically plausible, provided the first upload for 222 of these, and was the sole contributor for 75. Among Gravitational-Wave Transient Catalog events with a false-alarm rate below one per year, 88% were identified as significant in low latency and promoted for expert vetting and public dissemination. The low-latency astrophysical classifications agreed with the final catalog classifications for 93% of the events considered.