From the course: AI Pricing and ROI: A Technical Breakdown

Unlock the full course today

Join today to access over 25,200 courses taught by industry experts.

Ongoing costs for AI APIs

Ongoing costs for AI APIs

From the course: AI Pricing and ROI: A Technical Breakdown

Start my 1-month free trial Buy for my team

Ongoing costs for AI APIs

“

- [Instructor] Let's talk about the ongoing costs of using an AI API. There are three main ways to consume an AI API. The first one is managed hosting. The second is per API consumption. And the third one is a defined plan. Managed hosting is more traditional from a cloud perspective, where you pay a vendor to host a model and pay by the hour. Even though it's managed hosting, you can still connect easily via API. These hosting models are good if you have consistent workloads. Now as an example, between two vendors on the left here, we have AWS Bedrock serving pre-trained Llama with both one month and six-month commitments. And with Together.AI, you pay by model size per hour. Our second option is per API consumption pricing or serverless pricing. Here we pay by API call with some additional parameters like tokens that are used. We connected these models via an API, so it's still easy to use. And these are great for…

Contents