Model Routing
Opus where reasoning matters, Sonnet where speed and cost matter.
Pilot automatically routes each phase to the right model. Rather than always using the most powerful (and most expensive) model, it applies reasoning where reasoning has the highest impact — and uses fast, cost-effective execution where a clear spec makes quality predictable.
Routing Table
| Phase | Model | Rationale |
|---|---|---|
| Planning | Opus | Exploring your codebase, designing architecture, and writing the spec requires deep reasoning. A good plan is the foundation — invest here. |
| Plan Verification | Sonnet | The plan-reviewer sub-agent validates completeness and challenges assumptions on every feature spec. (enabled by default — disable in Console Settings → Reviewers) |
| Implementation | Sonnet | With a solid plan, writing code is straightforward. Sonnet is fast, cost-effective, and produces high-quality code when guided by a clear spec and strong hooks. |
| Code Verification | Sonnet | The unified spec-reviewer agent handles deep code review (compliance + quality + goal). The orchestrator runs mechanical checks and applies fixes efficiently. (enabled by default — disable in Console Settings → Reviewers) |
The Insight
- Implementation is the easy part when the plan is good and verification is thorough
- Pilot invests reasoning power (Opus) where it has the highest impact: planning
- Sonnet handles implementation and verification — guided by a solid plan and structured review agents
- The result: better output at lower cost than running Opus everywhere
Fully configurable
Configure via the Pilot Shell Console Settings tab (localhost:41777/#/settings). Choose between Sonnet 4.6 and Opus 4.6 for the main session, each command, and each sub-agent independently. Context window size (200K or 1M) is auto-detected from Claude Code based on your subscription plan — no manual configuration needed.