feat(core,runtime,desktop): add Fast mode for OpenAI relays

[hbw00111]

Summary

Adds an optional per-model Fast mode to custom OpenAI relay profiles. The setting maps to OpenAI service_tier: "fast" and remains independent from reasoning strength.

Refs #3300

Verification

• Core and runtime builds passed.
• Focused relay profile and provider-options tests passed.
• Desktop typecheck remains blocked by existing unrelated workspace drift.

AI use

• No generative tool made a substantive contribution
• Generative tooling made a substantive contribution

Tool(s) and scope: Codex performed codebase research, implementation, test authoring, and verification.

Checklist

• Tests cover the change and fail without it
• Lint, format, typecheck and affected suites pass locally

Does this PR entail a change in behavior?

• Yes — described under Summary above
• No

[hbw00111]

Fast mode is being folded back into PR #3300 so the approved feature stays in one review.