Cannot get LLM usage tokens/cost for each task in Guardrails (Jailbreak, Hallucination, Custom Prompt Check...)

[thuanng-a11y]

client = GuardrailsAsyncOpenAI(
            config=JsonString(json.dumps(guardrails_config))
        )
response = await client.chat.completions.create(
            model=os.getenv("OPENAI_MODEL_NAME", "gpt-4.1-mini"),
            messages=[{"role": "user", "content": req.input}],
            suppress_tripwire=True
)

I've checked the response object, it doesn't contain the LLM usage cost for each task of Guardrails (although the guardrails already checked and the tripwire triggered successfully). It just show the tokens used for LLM response.

How can I get that value? Or it has not implemented yet is this version?

Thanks.

[gabor-openai]

@thuanng-a11y thanks for flagging this. What cost information is returned to you when you use just the regular OpenAI client?

[thuanng-a11y]

@thuanng-a11y thanks for flagging this. What cost information is returned to you when you use just the regular OpenAI client?

Currently, i expect that it can return to me the input tokens and completion tokens of each task of Guardrails (e.g., when I use Custom Prompt Check with longer prompt, it will take me more tokens).

But now the library just return me the cost of for the LLM response. Here's the raw value I have copied from the output:

{
    "completion_tokens": 40,
    "prompt_tokens": 15,
    "total_tokens": 55,
    "completion_tokens_details": {
      "accepted_prediction_tokens": 0,
      "audio_tokens": 0,
      "reasoning_tokens": 0,
      "rejected_prediction_tokens": 0
    },
    "prompt_tokens_details": {
      "audio_tokens": 0,
      "cached_tokens": 0
}