Model Family

A quick-start guide to picking a first Claude model based on task and budget.

Explains the tiering logic behind Fable, Opus, Sonnet, and Haiku: power, speed, and cost.

Details Opus 4.8's 1M-token default context and where it fits versus Fable and Sonnet.

Explains why Haiku 4.5's low cost and 200K context suit high-volume, latency-sensitive tasks.

A decision list matching everyday tasks, from quick chat to deep analysis, to model tier.

Enumerates guidelines for matching model tier to task complexity, latency needs, and budget.