feat(agents): add GPT-6 and Opus 5.5 support

Merged
HA
2mo ago
+91-23

What & why

Add support for GPT-6 Sol, GPT-6 Luna, and Claude Opus 5.5.

  • Add GPT-6 Sol and GPT-6 Luna to the Superset Chat and Codex workspace model selectors.
  • Add Claude Opus 5.5 to the Superset Chat and Claude workspace model selectors.
  • Pass the selected model ID to the corresponding CLI when spawning a workspace.
  • Track API-equivalent input, output, cache-read, and cache-write costs for all three models.
  • Expose the supported reasoning-effort levels for the new GPT-6 models.
  • Verify the model IDs, capabilities, and pricing against the current OpenAI and Anthropic documentation.

How I tested it

  • bun test packages/shared/src/agent-models.test.ts packages/host-service/src/trpc/router/usage/history/pricing.test.ts
    • 67 passed, 0 failed
  • bun run lint
  • bun run typecheck
    • 42 tasks passed
  • Manually verified GPT-6 Sol and GPT-6 Luna in the Codex model dropdown.
  • Manually verified Opus 5.5 in the Claude model dropdown.
  • Manually verified GPT-6 Sol in Usage with an estimated cost.

Screenshots

GPT-6 Sol and GPT-6 Luna in the Codex model dropdown Claude Opus 5.5 in the Claude model dropdown GPT-6 Sol usage tracking

Checklist

  • PR title follows conventional commits (type(scope): subject)
  • bun run lint and bun run typecheck pass
  • Allow edits from maintainers

Summary by cubic

Adds GPT-6 Sol, GPT-6 Luna, and Claude Opus 5.5 to the Superset Chat and agent model selectors.

  • Passes the selected model ID to the Codex or Claude CLI when spawning a workspace.
  • Tracks API-equivalent input, output, cache-read, and cache-write costs for all three models.
  • Exposes the supported reasoning-effort levels for the GPT-6 models; Sol supports ultra, Luna tops out at max.
  • Bumps the pricing table updated date to 2026-09-22.