Track input and output tokens for each LLM call and display estimated API cost per run. Show per-iteration token counts in logs and the final summary. Include a cumulative cost report. This helps users budget their API spend and identify expensive tasks. The Anthropic API response already includes usage metadata.
Track input and output tokens for each LLM call and display estimated API cost per run. Show per-iteration token counts in logs and the final summary. Include a cumulative cost report. This helps users budget their API spend and identify expensive tasks. The Anthropic API response already includes usage metadata.