Show quick token usage statistics
Read the Claude stats from ~/.claude/stats-cache.json and display a summary including: 1. **Total token usage** - Input and output tokens combined 2. **By model breakdown** - Show usage for each model (Opus, Sonnet, Haiku) 3. **Cache efficiency** - Percentage of tokens served from cache 4. **Estimated cost** - Calculate cost using these rates per 1M tokens: - Opus: $15 input, $75 output - Sonnet: $3 input, $15 output - Haiku: $0.80 input, $4 output Format the output in a clean, readable table format.