Understands intent automatically
Classifies prompts for writing, coding, research, analysis, or creative tasks, then selects the model that consistently scores highest for that domain.
Feature
Smart Routing picks the right model every time
Auto Mode weighs speed, cost, and quality so your team gets the best answer without juggling providers — and you can still take control any time.
How Auto Mode keeps teams productive
Balances speed, quality & cost
Routes lightweight prompts to fast, efficient models like GPT-4o-mini or Claude Haiku, while reserving GPT-5 Pro, Claude Opus, or Gemini 2.5 Pro for complex work.
Keeps thinking when it helps
Automatically enables extended reasoning on models that benefit from it. For projects requiring quick output, it sticks to low-latency responses.
Fails over gracefully
If a provider experiences delays, Auto Mode retries on a backup model so chats keep flowing without manual intervention.
Control when you want it
Auto Mode handles the heavy lifting, but the moment you need a specific provider, manual controls are one click away — balanced with workspace guardrails so everyone stays on policy.
Personal routing preferences
Users can set personal defaults — prefer Claude for writing, Gemini for analysis, or Groq for speed. Auto Mode honors preferences while maintaining quality floors.
Workspace guardrails
Admins define approved providers and cap usage of premium models. Auto Mode complies with workspace policies automatically.
Model insights
Usage analytics show how often Auto Mode selects each provider, so teams learn which models perform best for specific departments.
Prefer manual control? Pin your favorite model in Chat or browse the Model Directory.
Frequently asked questions
How does Auto Mode decide which model to use?
Auto Mode analyzes each prompt, looking at task type, previous responses, and past success rates. It maps the request to the model with the highest quality and success metrics for that domain while weighing speed and cost constraints.
Can I override Auto Mode mid-conversation?
Yes. You can switch models at any time. When you do, Auto Mode remembers and factors your preference into future decisions for similar requests.
Does Auto Mode support new model releases?
Absolutely. When new models launch, we benchmark them and update our routing heuristics. Auto Mode starts using successful newcomers while phasing out underperformers.
What if we need to restrict certain providers?
Workspace admins can toggle providers on or off, set budget caps, and define routing policies by team. Auto Mode follows those constraints automatically.