feat(ai): use Luna as the default conversation model

[izadoesdev]

Use openai/gpt-5.6-luna as the Balanced conversation default for the dashboard, Slack, and shared agent calls without an override. Public MCP clients still choose their own model. Other tiers and separately configured Insights models retain their selections.

The final diff is one line. Luna pricing and fresh/cached cost coverage already landed in #788; the shared helper from #782 explicitly preserves Luna low/medium/high OpenAI reasoning options. Both PRs and query discovery #783 are now in this branch's staging base.

Validation: fresh root lint, all 33 type/build tasks, all 27 workspace test tasks, 56 native configuration/execution/billing/telemetry tests (605 assertions), and four shared cost tests pass. Independent review confirms the exact one-line diff and compatible provider options. Existing flat-rate accounting still omits long-context and service-tier surcharges.

Earlier live-model evidence: a 64-conversation synthetic evaluation favored Luna, with 18/22 grounded numeric answers versus Terra's 16/22 and 87.1% lower returned cost across 19 fully billed matched data pairs. That diagnostic evaluation used the conversation/discovery repairs now on staging; it is not a general quality or production-cost benchmark. No new live customer charge or production promotion was performed by this change.

AI-assisted implementation and verification under the maintainer workflow. Final-head CI and configured review are requested before staging merge.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
databuddy-status	Ready	Preview	Sep 14, 2026 12:54pm UTC

2 Skipped Deployments

Project	Deployment	Actions	Updated
dashboard	Skipped		Sep 14, 2026 12:54pm UTC
documentation	Skipped		Sep 14, 2026 12:54pm UTC

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Repository UI

Review profile: ASSERTIVE

Plan: Advanced

Run ID: a4784663-18d5-4c22-ac8f-e753f6423502

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[unkey-deploy]

The latest updates on your projects. Learn more about Unkey Deploy

Name	Status	Preview	Inspect	Updated (UTC)
links (preview)	Ready	Visit Preview	Inspect	Sep 14, 2026 12:54pm

[izadoesdev]

@greptileai Please review final head 8cbebfb before staging merge. The net change is one line: Balanced conversation model Terra to Luna. Staging already includes Luna pricing, native cost tests, and explicit Luna reasoning support from #788/#782, plus query discovery #783; duplicated additions have been removed. Fresh root lint, 33 type/build tasks, 27 test tasks, 56 native execution/config/billing/telemetry tests and four cost tests pass. Please verify the default paths and provider compatibility.

[greptile-apps]

Greptile Summary

Updates the shared Balanced conversation model from Terra to Luna.

• Changes modelNames.balanced to openai/gpt-5.6-luna.
• Leaves the Tiny, Quick, and Deep model selections unchanged.

Confidence Score: 5/5

The one-line model-default change appears safe to merge, with no actionable issues identified.

The final head contains only the intended Balanced model substitution, no additional changes have landed since the previous review SHA, and no repository-rule or security issues remain.

Important Files Changed

Filename	Overview
packages/ai/src/ai/config/models.ts	Changes the shared Balanced model identifier from Terra to Luna without altering other model tiers.

_{Reviews (2): Last reviewed commit: "feat(ai): use Luna as the default conver..." | Re-trigger Greptile}

[izadoesdev]

@greptileai Please retry the review for final head 8cbebfb. The previous check completed as NEUTRAL with “Review was cancelled” and produced no review. This PR is now ready and its net diff is one model-default line; the required pricing and Luna reasoning support are already on staging. Please provide a completed review of this head before staging merge.