Skip to main content

Pricing

Meter the runtime.
Not the ambition.

Sonzai pricing stays close to what is actually running: model usage, active agents, and any team support you choose to add.

Runtime

Provider rate + 25%

Use the model provider you already trust. Sonzai meters the orchestration, memory writes, context assembly, and retrieval layer around it.

  • OpenAI, Anthropic, Google, and other provider paths
  • Usage visibility by agent, user, and deployment
  • Works for REST, SDK, and MCP integrations

Active agents

$0.50 / agent / month

Only agents that are actively used during the month count. Dormant experiments and test agents do not create surprise bills.

  • Unlimited users per active agent
  • No seat pricing for customer-facing deployments
  • Simple monthly usage export for finance reviews

The team

Custom

For companies that want Sonzai to help design, build, or operate the first AI employee alongside their engineers.

  • Architecture and integration support
  • Private deployment and SLA conversations
  • Custom implementation scopes when the use case needs it
Contact us

Example monthly bill.

A small production pilot can stay cheap while still carrying real memory, tools, and workflow state.

2 active agents$1.00
10M tokens via GPT-4o-mini plus markup$7.50
Estimated monthly total$8.50

No hidden platform puzzle.

If you need procurement, a private deployment, or help scoping the first AI employee, contact the team and we will price that work separately from runtime usage.