Claude Model Selection and Tradeoffs

Audience: professionals who use Claude at work, no coding or API knowledge needed. Verified = checked against Anthropic's documentation or help center on 2026-10-16. Design guidance = widely taught practice, not a documented rule.

The model families (verified from the API models overview; version-sensitive)

All current models handle text and images. Which models you can choose in the Claude app depends on your plan and rollout (and your organisation's settings), so check the model picker you actually have. The names change with each release, so do not memorise them; learn the pattern.

The pattern: three trade-offs (design guidance)

Bigger models are generally slower and use more of your plan's usage, so do not default to the biggest for simple tasks.

How to decide

1. Start with a capable default and see if the result is good enough.
2. If it is good enough but slow or costly, try a smaller tier.
3. If it is not good enough (shallow, misses steps), try the next tier up, and improve your prompt (D1.1).
4. Judge by the whole job: cost per completed, correct task, not just per message (Anthropic's guidance for choosing models, verified).
5. Use the bigger model for the hard step and the smaller one for routine steps.

Usage limits (verified)

Heavy features use limits faster: Research sessions "can use up your limits faster" because they retrieve many sources. Running several threads in Projects also uses usage faster.

Exam traps

• Always the largest model.
• Always the smallest model for complex analysis.
• Memorising version names that will change.
• Switching models instead of fixing a vague prompt.
• Ignoring plan limits.