The AI model GPT-5.6 has made a significant leap forward in the study of prime number distribution, breaking the record for the largest prime gaps previously set by Terence Tao's team. Leveraging an innovative construction technique known as "tilted residue classes," the model not only advances the field but also translates proofs into Lean language code for rigorous formal verification, thereby pioneering a new research paradigm. Prior to this, GPT-5.6 had already refined related conclusions drawn by Yitang Zhang, addressing two fundamental questions in prime number distribution. The research documents related to these achievements have been made publicly available, highlighting AI's growing role as a valuable collaborator in mathematical research.
