Frontier model
A frontier model is one of the most capable general-purpose AI models available at a given time, usually built by a handful of well-funded labs using enormous computing power. Governments single them out because new abilities, and new risks, tend to appear there first.
In one line, for a 12-year-old
Frontier models are the newest, most powerful AIs — the ones at the edge of what's possible right now.
An example
The newest top models from Anthropic, Google DeepMind, OpenAI and a few others are usually called frontier models when they launch.
Why it matters to people
Because so few organisations can build them, decisions about their safety testing and release affect millions of people. Canada, the UK, the US and others have set up AI safety institutes partly to study these models.