Skip to content

Workflow Modes & Model Tiers

GloriaMundo uses workflow modes to decide which execution tier a workflow should use. You see these modes in two places.

First, Settings > AI Models > Mode sets your account default. New workflows use that setting automatically unless you later override them. Second, the Workflows dashboard lets each saved workflow stay on Default or switch to a specific mode.

That means there are four account-level modes, plus one workflow-level inheritance option:

Mode Scope What it means
Default Workflow override only Use the account default from Settings
Economy Account default or workflow override Lowest-cost execution tier
Balanced Account default or workflow override General-purpose tier for most workflows
Ultimate Account default or workflow override Highest-capability tier for more demanding workflows
Custom Account default or workflow override Use your own model selections

How mode selection works

For new builds, the builder no longer asks you to choose a mode before you start. It uses your account default automatically. If you have never changed this in Settings, the default is Balanced.

After you save a workflow, you can change its mode from the badge on the Workflows dashboard:

  • Leave it on Default to inherit the account default
  • Override it with Economy, Balanced, Ultimate, or Custom

This keeps the build flow simpler while still giving you per-workflow control once a workflow exists.

Economy

Choose Economy when you want the lowest-cost execution tier. This is the best fit for straightforward automations where speed and cost matter more than depth.

Use it for:

  • Simple data movement
  • Short summaries or straightforward formatting
  • Workflows you expect to run frequently and cheaply

The trade-off is that more complex workflows may need more manual refinement or a higher mode.

Balanced

Balanced is the default account mode and the best starting point for most users. It is the general-purpose tier for workflows that connect several steps, transform data, and generate user-facing output.

Use it for:

  • Day-to-day business workflows
  • Multi-step automations across connected services
  • Content, summaries, and analysis that need a reliable first pass

If you are unsure which mode to choose, start here.

Ultimate

Choose Ultimate when the workflow is more demanding and you want the highest-capability execution tier available in the current product.

Use it for:

  • More complex workflows with many moving parts
  • Higher-stakes workflows where quality matters more than speed
  • Runs that benefit from the most thorough available model tier

Ultimate is slower and more expensive than Economy or Balanced, so it is usually better as a deliberate choice than a blanket default.

Custom

Choose Custom when you want to control the model assignments yourself. In Settings, GloriaMundo exposes the execution roles you can customise directly, such as the models used for solving steps, user-facing writing, and related execution behaviour.

Custom is useful when:

  • Your team has a specific preferred model mix
  • One workflow family needs different trade-offs from the rest of the account
  • You want tighter control over the models used at execution time

If you select Custom for a workflow, that workflow uses your current custom selections instead of the standard tier bundle.

Default

Default only appears as a workflow-level option on the Workflows dashboard. It is not a separate model tier. It simply tells the workflow to inherit the account default you have already chosen in Settings.

Use Default when you want one central account-level setting and do not need to manage each workflow separately.

Per-step model override

Modes control the workflow-level execution tier. If one LLM step needs special treatment, you can still override that step in the editor without changing the rest of the workflow.

Use per-step overrides sparingly. They are most useful when one step is unusually important and the rest of the workflow can stay on a lower-cost mode.

Cost impact

Higher modes usually cost more. Use Virtual Run before a live run so you can see the expected cost and decide whether the current mode is appropriate.

See also: