Implement xAI (Grok) Model Provider

[Furisto]

Overview

The xAI provider is not implemented. Only model definitions exist (`grok-code-fast-1`, `grok-4-0709`, `grok-3`, `grok-3-mini`), but there is no provider implementation to invoke these models.

Current State

• ✅ Model definitions exist for all supported Grok models
• ✅ Pricing and capabilities defined
• ❌ No XAIProvider struct implementation
• ❌ No InvokeModel method
• ❌ No message transformation
• ❌ No tool/function calling support
• ❌ No tests

What Needs to be Done

Implement a complete xAI provider following the established patterns from OpenAI and Gemini providers.

Implementation Requirements

1. Provider Structure

   • Create `XAIProvider` struct with xAI client
   • Implement proper initialization with API key validation
   • Add configuration options support

2. Core Functionality

   • Implement `InvokeModel` method with streaming support
   • Handle request/response transformation
   • Implement proper error handling and logging

3. Message Handling

   • Transform internal Message format to xAI API format
   • Support user, assistant, and tool messages
   • Handle system prompts appropriately

4. Tool/Function Calling

   • Transform tools to xAI format
   • Handle tool calling responses
   • Support multiple tool calls

5. Streaming Support

   • Implement streaming chunk handling
   • Accumulate streamed responses
   • Handle stream errors gracefully

6. Token Usage Tracking

   • Track input and output tokens
   • Calculate and report token metrics
   • Handle cache tokens if supported

Files to Modify/Create

• `backend/model/xai.go` - Implement XAIProvider
• `backend/model/xai_test.go` - Create comprehensive test file

Implementation Pattern

Follow the patterns established in:

• `backend/model/openai_completion.go`
• `backend/model/gemini.go`

References

• xAI API Documentation
• OpenAI Provider: `backend/model/openai_completion.go`
• Gemini Provider: `backend/model/gemini.go`
• Provider interface: `backend/model/provider.go`