Domain 3·12% of exam·24 concepts
Product and Model Selection
Pick the right Claude product feature and model tier for a task, balancing cost, speed, and quality while managing context and memory limits.
3.1
Select appropriate Claude product features (Projects, research mode, chat, artifacts)
- ccaof-3.1.1The Four Claude Entry Points
- ccaof-3.1.2Project Anatomy: Instructions, Knowledge, History
- ccaof-3.1.3The Project Worth-Building Test
- ccaof-3.1.4Conversation Isolation Inside a Project
- ccaof-3.1.5Choosing Research Mode vs Chat Web Search vs Artifacts
- ccaof-3.1.6Entry Point Selection Under a Recurring-Workflow Scenario
3.2
Differentiate between Claude model types (Haiku, Sonnet, Opus)
- ccaof-3.2.1The Haiku / Sonnet / Opus Tier Spectrum
- ccaof-3.2.2Haiku Task Profile: Structured, High-Volume Work
- ccaof-3.2.3Sonnet Task Profile: General Professional Work
- ccaof-3.2.4Opus Task Profile: Complex, High-Stakes Judgment
- ccaof-3.2.5Unified Model Tier Decision Logic
- ccaof-3.2.6Model Availability Is Plan-Dependent, Not Fixed
3.3
Align model selection with task requirements (cost, speed, quality)
- ccaof-3.3.1The Speed-Versus-Capability Tradeoff
- ccaof-3.3.2Cost as a Usage-Budget Consideration
- ccaof-3.3.3Volume Compounds the Tier Decision
- ccaof-3.3.4Matching Tier to Task Stakes, Not Habit
- ccaof-3.3.5Combined Entry-Point and Model Scenario Matching
- ccaof-3.3.6Avoiding Over-Engineering and Under-Resourcing
3.4
Understand and manage context limitations and memory considerations (when to restart, summarize, or persist)
- ccaof-3.4.1The Context Window as a Finite Budget
- ccaof-3.4.2Recognising Signs of Context Degradation
- ccaof-3.4.3Three Responses to a Degraded Session: Restart, Summarise, Persist
- ccaof-3.4.4Choosing Between Restart and Summarise
- ccaof-3.4.5Memory Persistence, Scope, and Curation
- ccaof-3.4.6Diagnosing a Long-Session Context Failure