Nvidia CEO Jensen Huang recently claimed that artificial general intelligence has been achieved, sparking intense debate over the definition and measurement of human-level machine intelligence in the industry.

Key points

• Nvidia CEO Jensen Huang suggested AGI is already here, though his definition relied on an AI's ability to build a $1 billion company.
• Google DeepMind researchers recently proposed a "Cognitive Taxonomy" to measure AI across 10 human faculties, including reasoning and social cognition.
• The ARC-AGI-3 benchmark, created by François Chollet, focuses on testing an AI's ability to learn new skills and solve abstract puzzles.
• OpenAI’s internal contract with Microsoft reportedly defines AGI as a system capable of generating $100 billion in profits.
• Current AI models, including GPT-5, have scored roughly 57% on cognitive frameworks, indicating they still fall short of well-educated human performance.

The lack of a standardized definition for AGI allows tech companies to use the term as a marketing tool to generate hype while obscuring the actual limitations of current models. Establishing rigorous, scientific benchmarks is essential for investors and the public to distinguish between genuine technological breakthroughs and corporate branding.