Costs & Monitoring

Viewing Costs

In-session:

> /cost

Shows:

Current session tokens
Estimated cost
Model used

Token Usage

Model	Input	Output
Claude Sonnet	Lower	Lower
Claude Opus	Higher	Higher
Claude Haiku	Lowest	Lowest

Cost Optimization

Reduce Token Usage

Use /compact to compress context
Clear conversation when done
Use specific prompts
Choose appropriate model

Model Selection

Task	Recommended
Simple queries	Haiku
General coding	Sonnet
Complex tasks	Opus

Monitoring

Session Tracking

The status bar shows:

Token count
Cost estimate
Model name

Usage Reports

Check Anthropic console for:

Daily usage
Cost breakdown
API calls

Setting Limits

In settings:

json

{
  "limits": {
    "maxTokensPerSession": 100000,
    "maxCostPerSession": 5.00
  }
}