Model Selector
The Model Selector dropdown sits below the chat input bar and controls which AI model processes your prompt.
Auto-Route (Default)
When set to Auto-Route, NorthSignal's Intelligent Router automatically picks the optimal model for each message. This is the recommended setting for most users.
Auto-Route considers:
- Prompt complexity and length
- Available provider keys
- Model strengths and pricing
Manual Model Selection
If you want to lock a conversation to a specific model:
- Click the Model Selector dropdown below the input bar.
- Browse the available models, grouped by provider.
- Select your preferred model.
Your choice persists for the current conversation. Start a new chat to reset to Auto-Route.
Available Models
The models visible in the dropdown depend on which API keys you've connected:
OpenAI (requires OpenAI key)
| Model | Description |
|---|---|
gpt-5 | Flagship model — fast, versatile, great for most tasks |
gpt-5-mini | Lightweight version for simple, rapid tasks |
Anthropic (requires Anthropic key)
| Model | Description |
|---|---|
Claude Sonnet 4.5 | Top-tier for coding, analysis, and structured output |
Claude Haiku 4.5 | Fast and affordable for lighter tasks |
Google (requires Google key)
| Model | Description |
|---|---|
Gemini 2.0 Pro | Massive context window for long documents |
Gemini 2.0 Flash | Fast version with extended context |
Reasoning Mode Toggle
Next to the model selector, you'll find a Reasoning Toggle switch. When enabled:
- The AI will "think" more deeply before responding.
- Responses may take longer but will be more thorough and accurate.
- Best for complex coding problems, math, or multi-step logic.
When disabled, the AI responds as quickly as possible — ideal for conversational use.