3.2 · Lesson 2 of 4
Model Tiers
What you need to know
Claude comes in tiers with different characters. Haiku is the fast, lightweight option for simple high-volume work. Sonnet is the balanced everyday choice. Opus is the most capable tier for complex reasoning and nuanced judgement. Learn the character of each tier rather than memorising version numbers.
Three tiers, three characters
Haiku · Fastest and lightest
Quick lookups, short classification, simple reformatting, lightweight drafting, high-volume work.
Sonnet · Balanced capability and speed
Most everyday writing, analysis, and day-to-day business work — the sensible default.
Opus · Most capable
Complex, ambiguous, or high-stakes reasoning where depth matters more than speed.
Fair comparisons
When you trial tiers, use identical prompts and attachments in fresh chats so context does not leak. Differences show up most on hard, ambiguous tasks. On easy reformatting, the tiers often look alike — which is why defaulting to Opus for everything is usually wasteful.
Worked example
Run the same easy task (rename a list of 20 items) and the same hard task (analyse a messy trade-off scenario and recommend a path) across available tiers. Score against a short rubric for the quality you actually need. The easy task often clears the bar on Haiku; the hard task is where Opus earns its keep.
Exam traps
Memorise exact version numbers instead of tier character
Names and versions change. The exam tests Haiku / Sonnet / Opus character and fit.
Bigger is automatically better
A stronger model cannot fix a vague prompt or missing context. Routine work rarely needs Opus.
Judge tiers by which answer is longest
Judge against the quality the task needs, not against word count.
Assume every plan exposes every tier the same way
Available models depend on your plan and the current product lineup.
Practice scenario
You need to rename and format a list of twenty items the same way every morning. Which tier character is the best default fit?
Build exercise
Compare Haiku, Sonnet, and Opus on the Same Tasks
3.2 · Intermediate · ~35 min · claude.ai
What you will practise
- The character of each tier: Haiku fast and light, Sonnet balanced, Opus strongest on hard reasoning.
- How to run a fair comparison with identical prompts.
- How to judge a result on the quality you actually need.
- Why the difference shows up on hard tasks, not easy ones.
Use the Domain 3 comparison exercise to build a short model-choice guide from your own runs.
Open Domain 3 exercisesPath: /learn/claude-associate/exercises/3-product-model-selection
Sources
- CCAO-F Exam Guide v1.0 — Domain 3 task statement 3.2 — Model Tiers
- preporato/claude-certification-guide — CCAO-F §5.1 — tier character pedagogy (rewritten)
- Amey-Thakur associate-foundations notes — model table — distilled into site voice
- docs/ccao-f-implementation-plan.md — Domain 3 lesson map