# CCDV-F Practice: Model Selection | Your AI Department > 10 free CCDV-F practice questions on model selection and optimization (16.8% of the exam), with study notes and explanations. Source: https://youraidept.com/network/ccdv-f-practice-questions/model-selection-and-optimization CCDV-F domain drill # CCDV-F practice questions: Model selection and optimization. 10 original questions on Model selection and optimization, which carries 16.8% of the Claude Certified Developer – Foundations exam per the published guide. Study notes first, then the questions with an explanation for each. When this domain feels solid, take [the timed CCDV-F practice exam](https://youraidept.com/network/ccdv-f-practice-exam) and check the domain breakdown. ## What to know about Model selection and optimization - The model family trade-off: the most capable model for hard reasoning, the mid-tier for most production work, the smallest for high-volume classification and routing. - Cost is tokens in plus tokens out; latency is dominated by output tokens. Optimize both by trimming prompts, capping max_tokens, and caching. - Extended thinking budgets: when deeper reasoning pays for itself and when it is wasted. - Temperature and sampling settings, and why deterministic tasks want low temperature. - Routing patterns: a small model classifies, a larger model handles the hard cases. ## Model selection and optimization questions All domains (10) Model selection and optimization (10) Model selection and optimization A pipeline extracts three fields from short emails, millions of times a month. Which model tier makes sense? - The most capable tier, since extraction errors at this scale compound quickly - A mid tier with extended thinking enabled to raise accuracy on ambiguous fields - The tier with the largest context window, so that long quoted email threads are not truncated before extraction - The smallest tier that passes your eval, since cost and latency dominate at this volume Check answer ### Show answer Answer: The smallest tier that passes your eval, since cost and latency dominate at this volume. Match the model to the task. Simple, high-volume extraction is where the small tier earns its keep; measure with an eval first. Thinking adds cost on every call, which is the wrong trade for short emails. Model selection and optimization You need the same input to produce the same extraction every time. Which setting helps most? - Add a longer system prompt that describes each field in detail - Set temperature near 1 so the model commits firmly to a single answer - Set temperature near 0 so sampling favors the most likely tokens - Enable prompt caching so the cached prefix yields the same output Check answer ### Show answer Answer: Set temperature near 0 so sampling favors the most likely tokens. Low temperature makes sampling less random. It is the first lever for deterministic-style tasks. Caching reuses the prefix computation but does not constrain sampling. Model selection and optimization Which is the best practice for choosing a model tier for a new feature? - Use the largest tier for anything customer-facing and the smallest tier for internal tooling and scripts - Pick the tier with the largest context window so prompts will not need trimming - Prototype on a capable tier, then test whether a smaller tier passes your eval at lower cost - Start on the smallest tier and move up only when users report errors Check answer ### Show answer Answer: Prototype on a capable tier, then test whether a smaller tier passes your eval at lower cost. Measure quality per tier; do not assume. Waiting for user complaints means shipping a regression to find it. Model selection and optimization Output tokens dominate latency for a summarization endpoint. Which lever helps most? - Enable prompt caching on the source document so the whole response is served faster - Move the request to the Message Batches API for higher throughput - Constrain the output length and format so the model writes fewer tokens - Add extended thinking so the model plans before writing the summary Check answer ### Show answer Answer: Constrain the output length and format so the model writes fewer tokens. Fewer output tokens means faster responses. Caching speeds up input processing, which is not the bottleneck here. Model selection and optimization What does max_tokens control? - The size of the prompt prefix that is eligible for caching - The maximum number of output tokens the model may generate - The number of tokens kept when the context window overflows - The total tokens allowed across input and output for the request Check answer ### Show answer Answer: The maximum number of output tokens the model may generate. max_tokens caps output only. Input is bounded separately by the model's context window. Model selection and optimization A classification endpoint with a fixed label set should use which sampling setting? - Low temperature so the same input maps to the same label - A high top_p so rare labels are not excluded - High temperature so the model explores less obvious labels - Default temperature with top_k raised to widen the candidate set Check answer ### Show answer Answer: Low temperature so the same input maps to the same label. Deterministic tasks want low temperature. Widening the candidate set with top_k or top_p pushes the other way. Model selection and optimization When does extended thinking justify its cost? - On every request, since thinking is cached and free after the first call - On short factual lookups where the answer must be verified twice - On streaming requests, where thinking fills time while text arrives - On hard multi-step problems where accuracy matters more than latency Check answer ### Show answer Answer: On hard multi-step problems where accuracy matters more than latency. Spend thinking budget where reasoning pays. Thinking tokens are billed as output on each request; they are not cached away. Model selection and optimization Your bill is dominated by a stable 20k-token system prompt sent on every call. First optimization? - Add a cache_control breakpoint to the system prompt so its tokens are read from cache - Move the prompt into the first user message so it is billed as one block - Switch to a smaller model tier, since input tokens cost less there - Send the requests through the Batch API to get the discounted rate Check answer ### Show answer Answer: Add a cache_control breakpoint to the system prompt so its tokens are read from cache. Caching a stable prefix cuts repeated input cost without changing the model or latency. Batching also discounts but changes delivery time and does not fit interactive calls. Model selection and optimization Which workload is a good fit for the Batch API? - An IDE assistant that answers as the developer types - A voice agent that needs turn-by-turn responses - Nightly enrichment of a million records with no latency requirement - A live customer chat reply that must appear on screen within a couple of seconds Check answer ### Show answer Answer: Nightly enrichment of a million records with no latency requirement. Batches trade latency for cost. The other three are interactive and cannot wait for a batch to complete. Model selection and optimization What is a reasonable way to compare two models for your task? - Run the same eval set through both and compare accuracy, cost, and latency - Compare context window sizes and choose the larger one - Ask each model to rate its own fitness for the task - Pick the newer model, since each release is documented as improving on the last Check answer ### Show answer Answer: Run the same eval set through both and compare accuracy, cost, and latency. Head-to-head evals on your own data decide; release dates and self-assessments do not. ## Next - [Timed CCDV-F practice exam](https://youraidept.com/network/ccdv-f-practice-exam): 53 questions in 120 minutes, scored against the 720 pass mark. - [All CCDV-F practice questions](https://youraidept.com/network/ccdv-f-practice-questions) with a domain filter. - [The CCDV-F exam page](https://youraidept.com/network/ccdv-f): format, weights, registration, study path. ### Other CCDV-F domains - [Applications and integration (20 questions)](https://youraidept.com/network/ccdv-f-practice-questions/applications-and-integration) - [Agents and workflows (9 questions)](https://youraidept.com/network/ccdv-f-practice-questions/agents-and-workflows) - [Prompt and context engineering (7 questions)](https://youraidept.com/network/ccdv-f-practice-questions/prompt-and-context-engineering) - [Tools and MCPs (6 questions)](https://youraidept.com/network/ccdv-f-practice-questions/tools-and-mcps) ## FAQ ### How much of the CCDV-F exam is model selection and optimization? + 16.8% of the marks, per the published CCDV-F exam guide. The full domain list with weights is on the CCDV-F exam page. ### Are these real CCDV-F exam questions? + No. They are 10 original scenario questions written from the blueprint for this domain. Reproducing exam content would violate the candidate agreement. ### How should I study model selection and optimization for CCDV-F? + Read the domain notes on this page, answer the questions with explanations hidden, then take the timed CCDV-F practice exam and check the per-domain breakdown for this domain. ## Who wrote this ### [Matei Olaru](https://youraidept.com/about/matei-olaru) CEO, Your AI Department Former public-company CEO, corporate attorney, and World Bank consultant. Runs the YAID Network and the partner relationship with Anthropic. ### [Michael Greenberg](https://youraidept.com/about/michael-greenberg) AI Product Lead, Your AI Department AI at Google, Meta, and Samsung. Six patents. Leads member support and the certification prep guidance. Your AI Department is a registered Claude Partner Network company. Our members sit these exams under our partner standing. Facts on this page were last checked on October 15, 2026. ## Ready to sit CCDV-F for real? Registration needs a Claude Partner Network organization. [YAID Network](https://youraidept.com/network) members register under our partner standing. $149 per quarter, exam fee separate. [Apply to join YAID ↗](https://youraidept.com/network/apply?ref=drill)[How registration works →](https://youraidept.com/network/register-for-claude-certification)