A routine task
“Classify this incoming request and identify the team.”
A lighter model route may be sufficient. The chosen route still needs the permitted context and an acceptable result.
AiAx Router
The right model for the work.
Choose the model and reasoning effort around the task, rather than asking people to choose every time.
Different work needs different routes.
A difficult analysis
“Compare the proposed approaches and explain the dependencies.”
More demanding work can justify a stronger model or higher reasoning effort, subject to the configured provider and budget limits.
A restricted task
“Use only the deployment approved for this information.”
Policy must constrain the eligible routes before a model is chosen. EU hosting and zero-retention requirements depend on the selected service and configuration.
Understand the task
Classify the work and its difficulty before selecting a model. The Router implementation considers quality, cost and available capacity.
Match the effort
Reasoning effort follows task difficulty. More demanding work can receive deeper reasoning, while simpler tasks use a lighter route.
Respect the deployment
Sensitivity and permitted providers must be defined by the workspace deployment policy. The local Router implementation reviewed here does not by itself prove enforcement of those policies in every workspace.
Why choose open-weight?
Explore the benefits, deployment choices and trade-offs on the dedicated OpenWeight page.
Discover AiAx OpenWeight →