Model selection - the question platform teams ask first

A platform team will ask which model their engineers should use on day one, and "it depends" is the answer that loses the room. The reciting-features version of this question is everywhere; the depth that distinguishes a real operator is knowing the taxonomy, the cost loop and what Auto is actually doing under the hood.

There are three model categories, and the cue an engineer can see in the picker is the brain icon. A standard model (no brain icon) thinks once and responds - fastest and cheapest. A thinking/reasoning model (brain icon) runs chain-of-thought - slower and consuming far more tokens, so reserve it for deep reasoning, planning or when a standard model has hit a wall. Max Mode greatly enlarges the context window for jobs like auditing a very large codebase - but a bigger context degrades accuracy and costs more, so it's a selective tool, not a default.