A mobile app team is adding an FAQ-answering feature that must respond within a few hundred milliseconds on a limited compute budget, and the questions it needs to answer are narrow and repetitive, drawn from a small, well-defined set of app features. The Foundry model catalog offers both a large flagship model known for handling complex, open-ended reasoning and a smaller, lightweight model optimized for fast, low-cost inference on simpler tasks. Which model is the more appropriate choice for this FAQ feature?
Select an answer to reveal the explanation.
Short Explanation
This comes down to matching the tool to the job rather than always reaching for the biggest one available. The questions here are narrow, predictable and drawn from a small set of app features, and the response has to come back fast on a tight budget, which is exactly the situation a smaller, efficient model is built for. Bigger is not automatically better here, since the flagship model's extra reasoning power is meant for open-ended, complex problems, and paying its latency and cost for a repetitive FAQ lookup would not buy any real improvement in answer quality. There is also no rule saying only the largest models are allowed to power something running on a phone, that idea does not reflect how model selection actually works. And turning off safety filtering has nothing to do with making up for a smaller model's capability, those are two completely separate concerns, and filtering should stay on regardless of which model handles the answers.
Full Explanation
The correct answer is C. The FAQ feature has two defining constraints, a tight latency budget and a narrow, repetitive question set, and the lightweight model is built precisely for that profile, offering fast, low-cost inference well-suited to simpler, well-scoped tasks, so it can meet the response-time requirement without paying for reasoning capability the task does not need. Option A is incorrect because greater capability does not translate into better performance on every task; for narrow, repetitive questions the flagship model's added reasoning strength brings no meaningful accuracy benefit while likely costing more latency and money, which works against the stated constraints. Option B is incorrect because there is no rule tying flagship-only models to mobile-facing deployments; model choice depends on the task's requirements, not the client platform consuming the API. Option D is incorrect because model size and content safety filtering are unrelated settings, disabling filtering does not compensate for anything about model capability and would only remove a protection the deployment should keep regardless of which model is chosen.