Preset

Harbour desk · ticket_resolved · version 3

Output and reasoning

Max output tokensReasoning effort

Require structured output

Harbour reply follows a cost-first policy.
A served response has to score at least 0.85 and finish inside 1,800 ms at p95.
Anything priced above $0.0200 a request is outside this policy.
Right-size is off, so Parezen keeps the model the app asked for.
Cascade is off, so there is no cheap-first attempt.
If the chosen model fails, fallbacks run in order: openai/gpt-4.1.
Output is capped at 800 tokens, with reasoning effort low.
The route is marked async, so eligible work can take the provider's batch price.
Exact-match response cache is off for this route.
The near-duplicate (semantic) cache runs only on routes you list as safe for it.
Automation is manual. Nothing here applies on its own.

This sample has no cascade escalations. When one happens, its cost is included, not hidden.
Exact-match cache is off, so responses are not reused.
The near-duplicate (semantic) cache is left out of this projection; it runs only on routes you list.