Skip to main content
idapt
HomeAI ModelsPricing
Sign inStart free
Blog

How the Auto-Router Picks Your Model

Richard Morel· Founder·July 16, 2026·Last reviewed July 16, 2026

Auto mode answers the question most people do not want to think about per message: which of 200+ models should take this one? Here is what it does, so you can trust it where it earns trust and pin a model where you know better.

What Auto does

When a chat is set to Auto, the router classifies the message and routes it to a model tier: routine prompts go to fast, affordable models; hard prompts go to stronger ones. Two flavors exist:

  • Auto Smart: the everyday default, biased toward speed and economy.
  • Auto Expert: biased toward capability for consistently hard work.

The reply shows which model answered, always: routing is never a secret.

What it optimizes

The router's job is fit, not favoritism: match the prompt's difficulty to the cheapest tier that handles it well. Simple lookups on a frontier model waste money; hard reasoning on a budget model wastes your time. The pre-send cost estimate reflects the routed model, so economics stay visible.

When to pin instead

Auto is a default, not a doctrine. Pin a specific model when:

  • Consistency matters: long-form writing mid-piece should not change voice.
  • You know the winner: your side-by-side testing crowned a model for this job.
  • The task is niche: a coding specialist for a long coding session beats re-classification per message.

Switching is per-message: pinning now does not pin forever, and context follows the switch.

Good to know

  • Agents can be set to Auto or pinned per agent; pinning suits specialists.
  • Provider preferences and blocks apply to routed choices like everything else.
  • Auto never routes to your local models implicitly; local routing is its own explicit setting (see local inference).

Run it for a real day: leave a general chat on Auto, watch which models answer, and pin only where you disagree with its calls. The choosing models help article covers the full mechanics.

GuideModelsAI

Found this helpful? Share it:

  • What Auto does
  • What it optimizes
  • When to pin instead
  • Good to know

Continue reading

  • Capability-Matched Cost Cutting: Pay for the Ceiling You Use

    Most AI spend goes to flagship models doing non-flagship work. How to match benchmark tier to task tier, verify the swap, and keep quality where it matters.

  • Bring Your Own Keys: BYOK in Practice

    How BYOK works in idapt: connect provider API keys, route matching models on your accounts, and mix BYOK with platform billing and local inference.

  • One Prompt, Five Models: Comparing Side by Side

    How to run the same prompt across several models in idapt with parallel tabs, when triangulation beats trusting one answer, and how to pick a keeper.

  • Generate Images Side by Side

    How to run one image brief across two models in idapt, which model fits which job, and how iteration works when generation lives inside your chats.

Explore idapt

  • Best of: the picks by job

    Data-backed model picks for coding, writing, research, and media.

  • One chat. Every model.

    GPT, Claude, Gemini, Grok, Llama: pick mid-conversation, cheapest-first.

  • Home
  • Pricing
  • AI Models
  • Image models
  • Voice models
  • Video models
  • Rankings
  • New models
  • Model status
  • Multi-Model Chat
  • Voice Mode
  • Voice HUD
  • Web Search
  • Image Generation
  • Video Generation
  • Audio Generation
  • Transcription
  • Drive
  • Secrets
  • Sharing
  • Workspaces
  • Tasks
  • Memory
  • Agents
  • Subagents
  • Automations
  • Skills
  • Code Execution
  • Computers
  • Computer Use
  • Cloud Computers
  • Local AI
  • AI Gateway
  • API & SDK
  • CLI
  • MCP
  • Tunnels
  • All features →
  • Blog
  • Use cases
  • Comparisons
  • Best of
  • Learn
  • Free tools
  • Prompts
  • Templates
  • Changelog
  • Help center
  • FAQ
  • Privacy
  • Compare all models
  • Support
  • Developers
  • Quickstarts
  • API reference
  • API pricing
  • CLI
  • MCP
  • Downloads
  • Desktop
  • Badges and embeds
© idapt[email protected]TermsPrivacy PolicyLegal noticeReport content
X (Twitter)