---

name: token-budgeting description: Best practices for managing token usage and budgets tags: [tokens, cost, budgeting, llm]

Token Budgeting Skill

Use this skill to manage and optimize LLM token usage.

Commands

• /token-summary - Get daily token usage summary
• /token-budget - Check budget status

Best Practices

1. Monitor daily usage: Check summary regularly to track costs
2. Set budgets: Configure limits to prevent unexpected costs
3. Optimize prompts: Shorter prompts = fewer tokens = lower cost
4. Cache responses: Avoid repeated identical queries

Cost Estimation

Approximate costs per 1K tokens (varies by model):

• GPT-4o: ~$0.03 input, ~$0.06 output
• GPT-4o-mini: ~$0.0015 input, ~$0.006 output
• Claude 3.5: ~$0.003 input, ~$0.015 output