4 ms·
It might be cheaper for you to call an API to run the inference instead of renting a machine. GPT-4-Turbo goes for $0.01 on 1k input and $0.03 for 1k output on
by TriangleEdge 3y ago
It might be cheaper for you to call an API to run the inference instead of renting a machine. GPT-4-Turbo goes for $0.01 on 1k input and $0.03 for 1k output on Azure. A x2gd.2xlarg instance on AWS has 8vCPU and 128GB of memory. It goes for ~200$ per month.