A company has fine-tuned a custom model using an existing large language model (LLM) from Amazon Bedrock. They need to deploy this model to production to serve a consistent, steady rate of requests per minute.

What is the most cost-effective solution to meet these requirements?