When you start a new chat in Eden, the default model is set to Auto. Instead of choosing a specific AI model yourself, Auto automatically selects the best model for each message you send — optimizing for quality, speed, and cost.

What Auto Does

Auto is a routing system that reads your message and instantly decides which AI model should handle it. Different tasks have different requirements — a quick question doesn't need the same model as a deep research task, and a simple edit doesn't need the same model as a complex content creation workflow.

Auto handles this decision for you so you don't have to think about which model to pick. It analyzes what you're asking for — the type of task, the complexity, and whether it involves actions in your workspace — and routes your message to the model best suited for the job.

How It Decides

Auto classifies your message based on several factors:

What kind of task is it? Auto distinguishes between different types of work: agentic tasks (actions in your workspace like organizing files, creating documents, or researching), writing and content creation, summarization, quick edits, factual questions, brainstorming, translation, and more.

How complex is it? Simple questions get routed to fast, lightweight models. Complex reasoning or high-stakes decisions get routed to more powerful (and more expensive) models. This means you're not burning premium credits on tasks that don't need them.

Is it agentic or conversational? Auto makes an important distinction: if the AI needs to do something in your workspace (save a file, organize a folder, search and compile results), it routes to a model optimized for multi-step tool use. If the output is just returned in the chat (a summary, an explanation, feedback), it routes to a model optimized for that kind of response.

Which Models Does Auto Route To?

Auto selects from the following models based on your message:

You won't see which specific model is selected for each message — Auto handles this transparently in the background. The goal is to give you the best result for each task without requiring you to understand the differences between models.

Why Use Auto?

It saves credits. By routing simple tasks to cheaper models and reserving expensive models for when they're genuinely needed, Auto stretches your credits significantly further than if you used a single premium model for everything.

It matches quality to task. A fast model that excels at summarization will often give you a better summary than a premium reasoning model, because it's optimized for exactly that kind of output.

It handles agentic work well. When you ask Eden to do something in your workspace — organize files, research a topic, create documents — Auto routes to models that are specifically optimized for multi-step tool use and workspace actions.

When to Switch Off Auto

Auto works well for most people, but you might want to manually select a model if:

• You have a specific preference for a particular model's writing style

• You're doing a long session of the same type of work and want consistency

• You want to use a premium model like Claude Opus 4.5 for an entire conversation

• You need to stay on the same model to maximize caching benefits (switching models mid-chat resets the cache)

To switch, just select a different model from the model picker in any chat. You can always switch back to Auto at any time.

Auto and Your Credits

Because Auto routes to the most cost-effective model for each task, your credits generally go further with Auto than with any single model. Lightweight tasks like quick questions, summaries, and simple edits get handled by inexpensive models, while your credits are preserved for the tasks that genuinely benefit from more powerful models.

For a detailed breakdown of how credits work with each model, see AI Credits and Usage.