How can I access usage tokens (including reasoning tokens) when using streaming responses?

API
, , , ,
1

I’m trying to calculate the cost of my API calls when using oai reasoning models with streaming enabled. Where can I find the token usage information in the response? Should I disable streaming to get this data?

2

just found the solution, you can add the parameter:

"stream_options": {
        "include_usage": True
    }

eg:

data = {
    "model": "o1-preview",
    "messages": [{"role": "developer", "content": "hello"}],
    "stream": True,
    "stream_options": {
        "include_usage": True
    }
}
1 Like