fix(models): enable Priority on Gemini 3.7 Flash

[steebchen]

This branch originally added Gemini 3.7 Flash, but #3592 landed the same model first. Merged main and dropped everything that duplicated it. What remains is a single correction to the merged entry.

The change

The google-ai-studio mapping for gemini-3.7-flash declares serviceTiers: ["flex"], with the comment "priority requests are silently downgraded to standard for this model on the Gemini Developer API". That is not what the API does — Priority is served, and declaring only flex means routing never offers Priority on the model at all.

-				// priority requests are silently downgraded to standard for this
-				// model on the Gemini Developer API, so only flex is declared
-				serviceTiers: ["flex"],
+				serviceTiers: ["flex", "priority"],

Everything else on the entry — the introductory pricing, reasoningEfforts, the Vertex mapping with Flex removed — is taken from main unchanged.

Evidence

Direct requests to the Gemini Developer API with the same top-level service_tier field the gateway sends, three runs of each tier:

requested	served (usageMetadata.serviceTier)
priority	priority (3/3)
flex	flex (3/3)

The likely origin of the original finding: AI Studio reports the served tier in the response body, not in a response header, so a header-based check shows nothing and looks like a downgrade.

End-to-end through the gateway, from the log rows the e2e run wrote:

model	requested	used	prompt tokens	input_cost
google-ai-studio/gemini-3.7-flash	priority	priority	10	1.35e-05
google-ai-studio/gemini-3.7-flash	flex	flex	10	3.75e-06
google-vertex/gemini-3.7-flash	priority	priority	10	1.35e-05

The tier is resolved as served, not downgraded, and billing applies the provider's declared multipliers: 10 x 0.75e-6 x 1.8 = 1.35e-05 for Priority and 10 x 0.75e-6 x 0.5 = 3.75e-06 for Flex.

Verification

• TEST_MODELS="google-ai-studio/gemini-3.7-flash,google-vertex/gemini-3.7-flash" FULL_MODE=true over the gateway e2e files: 29 files passed, 134 passed / 39 skipped, 0 failures, including both service tiers on each mapping.
• pnpm exec vitest run packages/models — 138 passed.
• pnpm format and pnpm build clean.

Also independently confirmed while working on this branch, and matching what main already has: Vertex rejects Flex for this model (400 INVALID_ARGUMENT: "Flex API is not supported for model: gemini-3.7-flash") while Priority succeeds, so the Vertex mapping is correct as merged.

🤖 Generated with Claude Code

Summary by CodeRabbit

• New Features
  • Expanded Gemini 3.7 Flash availability through Google AI Studio to support both Flex and Priority service tiers.
  • Continued support for multimodal inputs, streaming, web search, caching, tool use, structured outputs, configurable reasoning levels, and usage configuration details.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 94b2547c-0009-4418-9560-4d23ba8caed6

📥 Commits

Reviewing files that changed from the base of the PR and between 987749d and 1afb6a4.

📒 Files selected for processing (1)

• packages/models/src/models/google.ts

---

Walkthrough

The Google AI Studio configuration for gemini-3.7-flash now supports both flex and priority service tiers.

Changes

Gemini 3.7 Flash service tiers

Layer / File(s)	Summary
AI Studio service tier configuration packages/models/src/models/google.ts	Updates gemini-3.7-flash to declare flex and priority service tiers. Removes the flex-only downgrade comment.

Estimated code review effort: 1 (Trivial) | ~3 minutes

Mergeability Score: ⚪ Minimal · up to 1afb6

This localized model-catalogue update adds the Gemini 3.7 Flash mappings and service-tier behavior; no actionable merge-blocking risk remains after normal checks and review.

Possibly related PRs

• theopenco/llmgateway#3592: Modifies the gemini-3.7-flash provider configuration in the same model catalog.
• theopenco/llmgateway#3170: Updates Google model service-tier mappings in the same file.
• theopenco/llmgateway#3176: Updates Google model service-tier metadata in the same file.

Suggested reviewers: smakosh

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely describes the main change: enabling Priority service support for Gemini 3.7 Flash.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch add-gemini-3-7-flash

---

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.}