[Stephan1226]

Summary

• Add documented standard API-equivalent rates for GPT-6 Sol, GPT-6.1 Sol, and GPT-6 Luna, including cache-write rates and long-context pricing (OpenAI API pricing, GPT-6.1 Sol model pricing).
• Add the official Claude Sonnet 5.5 rate (Anthropic pricing, model ID). Claude Opus 5.5 was already present.
• Keep codex-auto-review unpriced because no authoritative per-token pricing semantics are documented.
• Keep displayed costs as API-equivalent estimates, not Codex subscription charges.

Type of change

• Bug fix
• New feature
• Refactor / cleanup
• Documentation
• Other:

Validation

• swift build: passed.
• swift test --filter ModelPricingTests: 12 tests passed.
• swift test --filter UsageCostTests: 12 tests passed; 1 optional local-log audit skipped.
• ./scripts/test-gate.sh: the full suite reports 1,384 tests, 16 skipped, and 1 failure. LimitProgressRenderingTests.testPaceMarkerStaysCloseToPaintedTrackHeight measured 10pt against an expected 7pt (±1pt).
• Confirmed the same rendering test failure on upstream main at 42df4e6 by running it in isolation.

Checklist

• swift build and swift test pass locally (the one full-suite failure also reproduces on upstream main; see Validation)
• PR title and description are written in English
• UI changes are described above (no UI changes)
• No copyrighted assets, secrets, or private tooling references are committed (see CONTRIBUTING)
• Tests were added or updated for this change