Usage & Billing
Pricing
Pricing is model-specific and depends on input, output, and cached-input token rates.
Pricing components
- Input token price per million tokens.
- Output token price per million tokens.
- Cached input price per million tokens (for models that support it).
Local CLI-backed models may be billed through external provider accounts rather than in-app metering.
Cost-control playbook
- Use stronger models for hard phases, then switch to cheaper models for repetitive cleanup.
- Avoid unnecessary giant context attachments.
- Use subagents for exploration instead of bloating a single main thread.