LiveFR

Reasoning model

Definition: A reasoning model spends extra compute thinking before answering, improving hard tasks in logic, math and code.

Instead of answering in one pass, it works the problem through internally. This added thinking raises latency and cost, but also reliability on difficult problems.

See also

← Full AI glossary · AI news