fix: update Google playground model list

[thisisryanswift]

Summary

• Add Gemma 4 model IDs to the Google provider list in the playground
• Remove deprecated/shutdown Google model IDs from the playground picker

Validation

• Validated model IDs against Gemini API docs
• Ran pnpm run type-check with Node 24.14.1

[Copilot]

Pull request overview

Updates the Google provider’s model picker options used by the chat playground to reflect currently supported Gemini/Gemma model IDs.

Changes:

• Removed deprecated Google Gemini model IDs from the static Google provider model list.
• Added Gemma 4 model IDs to the Google provider model list.

[samuv]

Thanks for the update @thisisryanswift! Quick question before we merge: have you actually run a chat against gemma-4-31b-it and gemma-4-26b-a4b-it end-to-end?

I checked @ai-sdk/google@3.0.64 (the latest published version, also what we have locally) and these IDs aren't in the GoogleGenerativeAIModelId union yet — only Gemma 3 variants are listed (gemma-3-1b-it, gemma-3-4b-it, gemma-3-12b-it, gemma-3-27b-it, gemma-3n-e2b-it, gemma-3n-e4b-it). The union ends with (string & {}) so it'll typecheck fine, but that's just the catch-all, it doesn't tell us the SDK has actually been updated to route Gemma 4 correctly.

Could you confirm you've smoke-tested both models from the playground? Just want to make sure they actually respond and aren't silently 404'ing against the Generative Language API endpoint the SDK is pointed at. I'll also try to verify on my side once I get access to a Google API key again.