feat: add Kimi K2 thinking model to VertexAI provider

[roomote]

This PR attempts to address Issue #9268. Feedback and guidance are welcome.

Changes

• Added moonshotai/kimi-k2-thinking-maas model to the VertexAI provider configuration
• Configured with 256K context window (262,144 tokens for both max and context)
• Set pricing: $0.60 per million input tokens, $2.50 per million output tokens
• Model description: "Kimi K2 Thinking Model with 256K context window."

Testing

• TypeScript compilation passes without errors
• Linting checks pass
• No existing tests found for vertex models specifically

Closes #9268

---

Important

Add moonshotai/kimi-k2-thinking-maas model to VertexAI provider with specific configuration and pricing.

• VertexAI Provider:
  • Added moonshotai/kimi-k2-thinking-maas model to vertexModels in vertex.ts.
  • Configured with 262,144 max tokens and context window.
  • Pricing set at $0.60 per million input tokens, $2.50 per million output tokens.
  • Model description: "Kimi K2 Thinking Model with 256K context window."
• Testing:
  • TypeScript compilation and linting checks pass.
  • No existing tests for vertex models.

^{This description was created by for df3074b. You can customize this summary. It will automatically update as commits are pushed.}

[roomote]

Rooviewer   See task on Roo Cloud

Review complete. The implementation correctly adds the Kimi K2 thinking model to VertexAI with proper configuration matching the issue requirements. No issues found.

_{Mention @roomote in a comment to request specific changes to this pull request or fix all unresolved issues.}