feat: support injecting custom LLM clients into Gemini and AnthropicLlm

[brucearctor]

Enable injecting pre-configured LLM clients into Gemini and AnthropicLlm models to support multi-agent systems with distinct configurations. Addresses #5027.

Testing Plan

I have verified these changes with the following:

Automated Tests

• New Unit Tests: Added which specifically mocks the and clients to verify they are correctly consumed by the cached properties.
• Regression Testing: Ran all 86 existing unit tests in and to ensure legacy behavior via environment variables is preserved.

Manual Verification

• Pydantic Validation: Verified model initialization with actual and objects to ensure no runtime Pydantic resolution errors ().

Verification Command

uv run pytest tests/unittests/models/test_custom_client.py
uv run pytest tests/unittests/models/test_google_llm.py tests/unittests/models/test_anthropic_llm.py

[adk-bot]

Response from ADK Triaging Agent

Hello @brucearctor, thank you for your contribution!

To help reviewers better understand and verify your changes, could you please add a testing plan section to your PR description? This should describe how you've tested the changes, including any manual end-to-end tests you've performed.

You can find more details in our contribution guidelines.

Thanks!