OpenAI reportedly developed a method to track their progress towards building artificial general intelligence (AGI), which surpasses human capabilities. According to Bloomberg, the company shared a new classification system with employees, and intends to release it to investors and others in the future. OpenAI currently considers itself at Level 1, where AI can engage in conversational interactions. They believe they are approaching Level 2, where systems can solve problems at a doctoral-level education. Level 3 refers to systems acting on a user's behalf for several days, while Level 4 signifies AI capable of developing innovations. The highest tier, Level 5, pertains to AI systems performing the work of an organization.

OpenAI emphasizes that this classification system is a work in progress and is subject to change. OpenAI's CEO Sam Altman and CTO Mira Murati have stated their belief that AGI will be achieved within the next decade. Nvidia CEO Jensen Huang suggests that if AGI is defined as the capability to pass any given test, it may arrive within five years. Elon Musk has even suggested that AI could surpass human intelligence by 2025 or 2026. Additionally, Musk predicts that the "total amount of sentient compute" will exceed that of all humans within five years.