feat(ai-gemini): support gemini-3.5-flash

[elyase]

Summary

Please add first-class gemini-3.5-flash support to @tanstack/ai-gemini.

Google appears to have released the new Gemini 3.5 Flash model, and Vercel AI SDK has already merged support for the model ID in its Google providers. It would be helpful for TanStack AI users to get the same type-safe model support in the Gemini adapter.

Reference

Vercel AI SDK implementation PR:

• feat(provider/google): add gemini-3.5-flash vercel/ai#15435

That PR added gemini-3.5-flash to:

• @ai-sdk/google
• google.interactions
• @ai-sdk/google-vertex
• AI Gateway model ID unions
• Google / Vertex docs and examples

TanStack AI places that may need updates

Based on a quick scan of this repo, this likely touches:

• packages/typescript/ai-gemini/src/model-meta.ts
• packages/typescript/ai-gemini/tests/model-meta.test.ts
• packages/typescript/ai-gemini/tests/chat-per-model-type-safety.test.ts
• packages/typescript/ai-gemini/tests/tools-per-model-type-safety.test.ts
• docs/adapters/gemini.md

Expected behavior

Users should be able to do something like:

import { chat } from "@tanstack/ai"
import { geminiText } from "@tanstack/ai-gemini"

const stream = chat({
  adapter: geminiText("gemini-3.5-flash"),
  messages: [{ role: "user", content: "Hello" }],
})

…and have the model ID accepted by the adapter/types with the appropriate Gemini model metadata and provider-option support.

Notes

I searched existing TanStack/ai issues for gemini-3.5-flash, Gemini 3.5 Flash, and related Gemini terms and did not find an existing issue for this exact model support request.

[AlemTuzlak]

@elyase added, will be included in the next release