Skip to main content
Mythos

Model Routing (OpenClaw)

Model routing in 📝OpenClaw is this library's name for how the platform chooses which AI model an agent uses and what it does when that model fails, as described in OpenClaw's documentation as of November 2026.

OpenClaw identifies models by provider/model references. Its documented selection order starts with the primary model in agents.defaults.model.primary, then tries the models in agents.defaults.model.fallbacks in order. Auth-profile rotation happens inside a provider before OpenClaw moves to the next fallback model.

The model failover documentation describes failure handling in two stages: auth-profile rotation within the current provider, then model fallback to the next listed model. When a failure explicitly identifies an unknown, missing, or retired model, OpenClaw skips other credentials and advances to the configured fallback; without one, it surfaces the error. Auth failures put a profile into cooldown, and a rate-limited profile enters cooldown only when the provider says how long to wait.

Related settings narrow or extend the choice. agents.defaults.modelPolicy.allow is an optional allowlist of model references. agents.defaults.imageModel is used only when the primary model cannot accept images, and agents.defaults.utilityModel is an optional lower-cost model for short internal tasks such as generated session titles. A per-agent agents.entries.*.model overrides the default, as covered under 📝Multi-Agent Routing.

The documentation's runtime configuration page lists "model routing" among its topics, where the models.* root governs provider catalog behavior and custom provider definitions.

Contexts

MythOS is a product of Cooperating Citizens of One, Inc.

TermsPrivacyRefundsContact

© 2026 Cooperating Citizens of One, Inc. · MythOS since 2017 · Not affiliated with Anthropic.