[Bug] Gemini requests failed with 404 error

Bug description
When using spring-ai-starter-model-google-genai with version 1.1.0-M4, all API requests using the GoogleGenAiChatModel fail with an HTTP 404 Not Found error.

Upon inspecting the exception, the URL being called is: https://generativelanguage.googleapis.com/v1beta/models//gemini-2.5-flash/:generateContent

There appears to be an extra forward slash (/) between /models and the model name (//gemini-2.5-flash). This malformed URL is likely causing the 404 error, as the Google API endpoint does not recognize this path.

Environment

• Spring AI Version: 1.1.0-M4
• Java Version: 21
• Spring Boot Version: 3.5.7
• Language: Kotlin
• Vector Store: None
• Chat/Embedding Model: Google Gemini (gemini-2.5-flash)

Steps to reproduce

1. Create a new Spring Boot project (v3.5.7) using Kotlin.
2. Add the spring-ai-starter-model-google-genai dependency at version 1.1.0-M4.
3. Configure application.properties with the specified model:

   spring.ai.google.genai.api-key=YOUR_API_KEY_HERE
   spring.ai.google.genai.chat.options.model="gemini-2.5-flash"
   

4. Inject the ChatClient (or ChatClient.Builder) into a service.
5. Attempt to make any chat call, for example: chatClient.call("This is a test prompt").
6. Observe the HTTP 404 error response from the Google API.

Expected behavior
The GoogleGenAiChatClient (or its underlying components) should construct the correct URL without the extra slash.

The expected URL should be:
https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash/:generateContent

The API call should succeed (assuming a valid API key) and return a ChatResponse.

Minimal Complete Reproducible example

@RestController
@RequestMapping("/api/chat")
class ChatController(
    private val chatClientBuilder: GoogleGenAiChatModel
) {

    @GetMapping("")
    fun chat(): Map<String, String?> {
        val res = chatClientBuilder.call("Tell me who Jack Sparrow was in 30 words")
        return mapOf("response" to res)
    }
}

Start at GoogleGenAiChatModel and the underlying Google GenAI request URL construction, then reproduce the failure with the provided Kotlin example. Verify that the generated URL has no extra slash between /models and gemini-2.5-flash, and confirm the request succeeds with a ChatResponse.