Best API Engine Model for Key Phrase Extraction from Blog Posts

We are using 3.5 Turbo and are starting to hit the 90K token per minute rate limit. We are using the API for key phrase extraction from blog posts (see prompt below). I am curious to know if there are other models like babbage-002 or davinci-002 that could handle this task. I know we need to test it ourselves, but maybe someone has experience here.

"role": "assistant",
            "content": "From the provided website content, extract up to 10 high-relevance key phrases that users are most likely to search for and can serve as effective anchor texts for internal linking. If there are fewer than 5 high-relevance terms, return only those. Adhere to these guidelines: \n1. Each phrase should have a minimum of two words and a maximum of five words. \n2. The phrases should be unique and capture the main themes of the content. \n3. Exclude any numbers from the phrases. \n4 Output the results in json. \n Here is the text: \n\n \