fix(model): refresh Gemini provider presets

[c121914yu]

Summary

• Add current Gemini embedding model IDs gemini-embedding-2 and gemini-embedding-001.
• Remove Gemini presets that official docs mark as shut down or reaching shutdown date: gemini-3.1-flash-lite-preview, gemini-1.5-flash, gemini-1.5-pro, and text-embedding-004.

Evidence

• https://ai.google.dev/gemini-api/docs/models
• https://ai.google.dev/gemini-api/docs/deprecations
• https://ai.google.dev/gemini-api/docs/embeddings

Validation

• node .codex/skills/model-provider-updater/scripts/model_provider_presets.mjs apply-plan --plan /private/tmp/gemini-provider-plan-20260525.json --dry-run
• node .codex/skills/model-provider-updater/scripts/model_provider_presets.mjs inventory --provider Gemini
• git diff --check
• node_modules/.bin/tsc --noEmit

bun run test -- modules/model/provider/Gemini found no matching test files. Repo-wide bun run test was attempted and failed on existing temp-worktree alias resolution plus Feishu network authentication (open.feishu.cn DNS), unrelated to this provider-only change.

[Copilot]

Pull request overview

Updates the built-in Gemini provider model preset inventory to reflect current Google Gemini model availability, keeping the /models inventory aligned with official docs.

Changes:

• Added current Gemini embedding presets: gemini-embedding-2 and gemini-embedding-001.
• Removed deprecated/shutdown (or near-shutdown) Gemini presets: gemini-3.1-flash-lite-preview, gemini-1.5-flash, gemini-1.5-pro, and text-embedding-004.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

✅ Build Successful - Preview Images for this PR:

registry.cn-hangzhou.aliyuncs.com/fastgpt/fastgpt-plugin-pr:5c4deb8a78acb6eea9c01cf1ca5dee289b2395b6

Changed packages:
None