frontier model

A frontier model is one of the most capable AI systems available at a given moment, the leading edge of what the technology can do. The term is usually reserved for the largest, most advanced models from a handful of well-resourced labs such as OpenAI, Anthropic and Google DeepMind.

Frontier models are expensive to build, requiring vast compute and data, which is why so few organisations produce them. They also sit at the centre of safety and policy debates, since the most powerful systems carry both the greatest promise and the greatest potential for misuse.

A recurring theme in the news is how quickly open-weight models are catching up to the frontier, narrowing a gap that once seemed to belong permanently to the biggest closed providers.