Model Configuration Intermediate

Cline supports a wide range of AI providers and models. This lesson covers how to configure each provider, compare models for different tasks, track costs, and use custom API endpoints.

Supported Providers

Cline works with all major AI providers and several local model solutions:

Configuring Each Provider

Anthropic (Direct)

1. Open Cline settings (gear icon in the Cline panel)
2. Select Anthropic as the API provider
3. Paste your Anthropic API key
4. Choose a model (Claude Sonnet 4 recommended for most tasks)

Provider:  Anthropic
API Key:   sk-ant-api03-xxxxxxxxxxxxx
Model:     claude-sonnet-4-20250514

OpenAI

1. Select OpenAI as the API provider
2. Paste your OpenAI API key
3. Choose a model (GPT-4o recommended)

Google Gemini

1. Select Google Gemini as the API provider
2. Paste your Google AI Studio API key
3. Choose a Gemini model

OpenRouter

1. Select OpenRouter as the API provider
2. Paste your OpenRouter API key
3. Browse and select from all available models

Local Models (Ollama)

1. Install Ollama on your machine
2. Pull a model: ollama pull deepseek-coder-v2
3. In Cline settings, select Ollama as the provider
4. The model list will auto-populate from your installed models

Local Models (LM Studio)

1. Install LM Studio
2. Download a model from the LM Studio library
3. Start the local server in LM Studio
4. In Cline settings, select LM Studio as the provider

Model Comparison for Cline Tasks

Different models excel at different types of tasks. Here is a practical comparison for common Cline operations:

Cost Tracking

Cline shows the running cost of each task in real time. The cost display appears at the bottom of the Cline panel and shows:

• Input tokens: How many tokens Cline sent to the API (your messages + file contents)
• Output tokens: How many tokens the model generated
• Estimated cost: Calculated based on the model's pricing
• Total session cost: Cumulative cost across all tasks in the session

Custom API Endpoints

Cline supports custom API endpoints for organizations that run their own AI infrastructure or use API proxies:

Provider:     OpenAI Compatible
Base URL:     https://your-proxy.company.com/v1
API Key:      your-internal-api-key
Model ID:     your-model-name

This is useful for:

• Corporate AI proxies and gateways
• Self-hosted models with OpenAI-compatible APIs
• Cost-tracking proxies that log API usage
• Regional API endpoints for data compliance

Switching Between Models

You can switch models at any time during a session. Common strategies include:

• Start cheap, escalate: Begin a task with a fast, cheap model (Haiku/Flash). If it struggles, switch to a more capable model (Sonnet/GPT-4o) mid-task.
• Model per task type: Use premium models for complex tasks and budget models for simple ones.
• Local for sensitive code: Switch to Ollama/LM Studio when working with proprietary code that shouldn't leave your machine.