fix: add Google models gemini-3.1-flash-lite

[github-actions]

fix: add Google models gemini-3.1-flash-lite

Closes #540

Source issue: #540

Summary

Field	Value
Provider	google
Primary model	gemini-3.1-flash-lite
Changed models	gemini-3.1-flash-lite
Added models	gemini-3.1-flash-lite
Updated models	None
Verification sources	1 2 3

Verified metadata

Model	Display name	Parent	Providers	Format	Flavor	Token limits	Pricing	Lifecycle
gemini-3.1-flash-lite	Gemini 3.1 Flash-Lite		google	google	chat	input=1048576, output=65536	in/out=0.25/1.5 per 1M; cache read=0.025 per 1M	multimodal=true; reasoning=true

Verification notes

No LLM verification step ran — model metadata was already complete in the issue.

sync_models vs proposed update

sync_models cross-check found differences. Official provider verification was used for the applied values, and sync_models discrepancies are listed below for review.

Model	Field	Proposed update	sync_models	sync_models source models
gemini-3.1-flash-lite	catalog entry	present	missing	None

[vercel]

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

Project	Deployment	Actions	Updated (UTC)
ai-proxy	Ready	Preview, Comment	May 12, 2026 5:10pm

[github-actions]

Proxy model verification

Verification completed successfully.

Passed: 1
Failed: 0

Model	Status
gemini-3.1-flash-lite	✅ Passed

View job details

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: f6614fe4db

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "Codex (@codex) review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "Codex (@codex) address that feedback".

[chatgpt-codex-connector]

Do not expose token budgets for Gemini 3.1 Flash-Lite

For Gemini 3.1 models, Google's thinking docs recommend thinkingLevel and describe thinkingBudget as the 2.5-series budget control; this proxy currently maps any reasoning_budget or disable request to thinkingConfig.thinkingBudget in openaiParamsToGeminiMessageParams. Marking this new Gemini 3.1 model as supporting reasoning_budget will cause UIs/clients that rely on the catalog to expose numeric budgets or reasoning_budget=0, which are not the right controls for 3.1 Flash-Lite and can produce ineffective or unexpected reasoning behavior.

Useful? React with 👍 / 👎.