We are currently using Azure as our hosting service, and OpenAI as our LLM provider. What we need is a cost analysis for switching to Google/Gemini, Amazon/Bedrock, Meta/Llama, and any other LLM services I am missing. We want to know the costs of the servers themselves (i.e., all the administrative setup) + the cost of the LLMs themselves (i.e., input and output tokens). We also want to know the costs (and technical complexity) of setting up our own servers and running an LLM (which one would work with that setup?). In short, we need a very detailed analysis of all situations because very soon we will have to reduce our operating costs to the extreme, and we need a report in language that us non-technical people can understand. e.g., X option is overall the best because Y, Z option is the cheapest but comes with these additional non-monetary costs, A option is the easiest to implement but has a cost of B.....etc.
Years: Any
Location: Anywhere
Requested on: 2024-10-16
Artificial intelligence, Deep learning, Open AI, Machine learning, RAG, PyTorch