In MLOps, what is a model registry used for?

Correct answer: Managing model versions, metadata, and deployment status

Option B is correct because a model registry in MLOps is a centralized repository that tracks model versions, associated metadata such as training parameters and performance metrics, and deployment status across environments like staging and production, enabling reproducibility and governance. Option A is incorrect because raw training data is stored in feature stores or data lakes, not a model registry. Option C is incorrect because automated code testing is handled by CI pipelines and unit test frameworks, not the model registry. Option D is incorrect because API latency monitoring is a function of observability and model monitoring tools, not the model registry.