AI and LLMs · Middle · Question 35 of 75
Why add a fallback provider or model for an AI feature?
- 1Providers require at least two models per app
- 2Fallbacks always give better answers than the primary
- 3An outage at one provider shouldn't take the feature down
Show the answer
Correct answer: An outage at one provider shouldn't take the feature down
Explanation
Fallbacks need their own evals, since a different model may format or behave differently.