Getting Token Usage as part of the Final Answer

[mattmarcin]

Is your feature request related to a problem? Please describe.
No

Describe the solution you'd like
For cost/usage tracking it would be nice to be able to get the used input and output token count or even actual API usage expense.

Is this not possible with the current options.
It appears it may be possible to hook into the Monitor to get the token usage data or wrap it into the agent's run.

Describe alternatives you've considered
I looked into modifying the MultStepAgent.run function to bring in the token counts from the monitor.

if stream:
    # The steps are returned as they are executed through a generator to iterate on.
    return self._run(task=self.task, images=images)
# Outputs are returned only at the end as a string. We only look at the last step
final_step = deque(self._run(task=self.task, images=images), maxlen=1)[0]

if include_tokens and final_step.final_answer is not None:
    token_counts = self.monitor.get_total_token_counts()
    return {
        "answer": final_step.final_answer,
        "token_counts": token_counts
    }
return final_step

Additional context
Being able to get usage/cost data from API calls would help from a business perspective to track cost of execution and aid in iterating on agents to lower costs.