Default Behavior
The platform automatically selects models optimized for:- Agentic reasoning and planning
- Tool use and function calling
- Following complex instructions
- Multi-step task execution
Available Models
Claude
Anthropic’s models, known for nuanced reasoning and safety.
GPT
OpenAI’s models, versatile and widely capable.
Gemini
Google’s models, strong at knowledge tasks.
Unified Interface
The platform provides a unified interface for all models:- Switch models even mid-conversation
- Same capabilities work across all models
- Consistent behavior and tool usage
- No need to learn different APIs
Advanced Configuration
Thinking Tokens
Thinking Tokens
Enable thinking tokens for greater reasoning capabilities. The model shows its reasoning process, which improves complex problem-solving at the cost of increased latency and cost.
Model Overrides
Model Overrides
Set a specific model for an agent that should always use a particular LLM, regardless of default settings.
Cost vs. Capability
Cost vs. Capability
Choose faster, cheaper models for simple tasks and more capable models for complex reasoning.
When to Change Models
| Scenario | Recommendation |
|---|---|
| Complex reasoning | Use models with thinking tokens |
| High volume, simple tasks | Use faster, cheaper models |
| Specific domain expertise | Test different models for best results |
| Cost optimization | Start capable, then optimize |
For most use cases, the default model selection works well. Only change models if you have specific requirements or are optimizing for cost.