Apart Research

Apart Research has announced an AI Verification Grand Challenge for security researchers, cryptographers, ML and GPU engineers and related specialists. Teams will tackle technical verification problems online, with selected finalists progressing to an expenses-paid final in Singapore. The program promises cash prizes, compute and API support, but has not published prize amounts, final dates or an application form.

Apart Research's fellowship supports promising participant-led AI safety projects that emerge from its research Sprints and Apart Studio. Selected teams receive 6 to 16 weeks of research management, compute and publication support, including travel to present their work. It is remote and part-time, with a minimum ten-hour weekly commitment; entry is by invitation and there is no stipend by default.

Apart Research runs recurring AI safety research hackathons open to participants worldwide, with research projects judged for cash prizes and selected teams invited to continue through the Apart Fellowship. Current and announced rounds cover swarm alignment, epistemics, agent attribution and frontier-AI threat modelling. Published cash prize pools are $4,000 for Swarm Alignment, $2,000 each for Epistemics and Agent Attribution, and $5,000 for Threat Intelligence and Threat Modelling.