Control Gemini request duration with a new timeout parameter

The Gemini model class now accepts a timeout parameter, giving teams explicit control over how long a request is allowed to run before being cancelled. This is particularly useful for production deployments where unbounded request durations can affect reliability and resource utilization.

Details**:**

• Set timeout (in seconds) directly on the Gemini model instance
• Applies to all request types made through the Gemini model class
• Falls back to the existing default behavior when not set; no migration required

See reference in docs.