AI Model Choice: Claude Haiku vs Sonnet for Creator Workloads

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-model-families-AI-picking-claude-haiku-vs-sonnet-r13a3-creators

1. A creator needs to extract product names from 500 customer emails. Which approach is most cost-effective?

   1. Use the most expensive model available for reliability
   2. Use Haiku because extraction is a well-scoped task that benefits from low latency
   3. Ask the AI which model to use for this task
   4. Always use Sonnet for data extraction to ensure accuracy

2. What is the primary advantage of Haiku over Sonnet for suitable tasks?

   1. Better handling of highly ambiguous prompts
   2. Lower latency and cost while maintaining quality for well-scoped tasks
   3. Guaranteed identical outputs across all versions
   4. Superior multi-step reasoning ability

3. A student builds a system that first attempts a query on a cheap model and only escalates to a more expensive model if confidence is low. What is this strategy called?

   1. Quality benchmarking
   2. Cascading
   3. Premium routing
   4. Parallel processing

4. A creator asks an AI which Claude model they should use for their YouTube video descriptions. The AI gives a strong recommendation. What does the lesson say about this?

   1. The AI recommendation should always be followed for safety
   2. The AI can accurately recommend the best model based on general information
   3. The AI cannot tell you which model is best without seeing your specific prompts
   4. The AI knows the current pricing of all models

5. Which type of task is best suited for Sonnet rather than Haiku?

   1. Writing a creative story that requires maintaining character consistency across multiple chapters
   2. Translating a simple sentence into Spanish
   3. Counting word frequency in a document
   4. Extracting names from a structured form

6. What does the lesson identify as a key risk of always choosing Sonnet 'to be safe'?

   1. Sonnet might produce lower quality outputs for simple tasks
   2. It burns budget unnecessarily on tasks where Haiku suffices
   3. Sonnet is always slower than Haiku
   4. Sonnet cannot handle well-scoped tasks

7. A creator wants to know if Haiku and Sonnet will produce identical outputs for the same prompt. What does the lesson say?

   1. OnlyHaiku guarantees consistent outputs
   2. Yes, both models guarantee identical outputs
   3. No, AI cannot guarantee identical behavior across model versions
   4. Only Sonnet guarantees consistent outputs

8. A creator is building a system to predict next quarter's API pricing for planning. What does the lesson say about this?

   1. AI pricing predictions are 100% reliable
   2. AI cannot predict next-quarter price changes from the vendor
   3. AI should be consulted for all pricing decisions
   4. AI can accurately predict pricing based on market trends

9. For a classification task that categorizes incoming support tickets into 5 fixed categories, which model would the lesson likely recommend as a first choice?

   1. Haiku, because classification into fixed categories is a well-scoped task
   2. Sonnet, because classification requires deep reasoning
   3. Both equally, because the task is simple
   4. Neither—classification requires human judgment

10. A creator reads that Sonnet 'reasons better' than Haiku. In which scenario would this matter most?

    1. Answering a customer email that could have multiple valid interpretations
    2. Generating a list of ingredients from a recipe image
    3. Extracting phone numbers from business cards
    4. Translating a single sentence to French

11. Before committing to a model for a production workflow, what does the lesson recommend?

    1. Use the most popular model among competitors
    2. Benchmark both models on your actual prompts
    3. Choose the cheapest model available
    4. Read vendor marketing materials

12. A creator notices their workflow sometimes uses Haiku and sometimes Sonnet depending on the query complexity. What key metric should they monitor to optimize costs?

    1. The model response length
    2. The percentage of queries escalated from Haiku to Sonnet
    3. User satisfaction scores
    4. The number of API calls per day

13. What does the lesson say about assuming you need to upgrade to Sonnet?

    1. There is no reason to ever use Sonnet
    2. Assume you need Sonnet for most tasks to be safe
    3. Always start with Sonnet and downgrade if unnecessary
    4. Measure quality on Haiku before assuming you need to upgrade

14. A creator wants to use AI to determine the best model for their workload by describing their tasks. Can AI reliably make this determination?

    1. Yes, AI can reliably recommend the best model from a description
    2. No, AI cannot tell you which model is best without seeing your prompts
    3. No, AI should never be consulted for model selection
    4. Yes, but only for simple tasks

15. Which scenario best illustrates the 'task fit' concept from the lesson?

    1. Using Haiku for straightforward extraction and Sonnet for complex reasoning tasks
    2. Ignoring model differences and using one model for everything
    3. Always using the most expensive model because it must be best
    4. Choosing models based on how catchy their names sound