Sume pricing has two parts. A plan sets what you can use and how many jobs run at once. Usage is billed from one prepaid wallet at each model's published USD rate, and video generation, the Sume Agent, Formats, and the API all draw from that same balance.
The tables below are generated from the same price list that Pricing and API pricing render, so they match those pages.
What plans are there?
| Plan | Price | Concurrent jobs | Includes |
|---|---|---|---|
| Free | $0 | 1 | No credit card required; Access to image generation models; Limited access on Sume Agent |
| Pro | $40 / month | 4 | Access to 100+ video generation models; Full access on Sume Agent; API + CLI + MCP access; Team wallet and on-demand billing |
| Startup | $120 / month | 8 | Shared team workspace |
| Scale | $400 / month | 20 | Highest self-serve concurrency; Priority support |
| Enterprise | Custom | Custom | Custom plans and volume pricing; Invoice / PO billing; Higher concurrency and controls; Priority support and account management |
How is usage billed?
Generation and API calls are billed at each model's published USD rate. Credits sit in one balance: there is no separate agent wallet. Teams fund a shared wallet with on-demand purchases from $10 to $1,000 from the dashboard's Billing & subscription page; larger organizations can ask for custom volume terms.
The public Developer API reads the wallet with GET /v1/balance and GET /v1/usage. Top-ups are a dashboard operation; there is no public top-up endpoint.
How are model rates set?
Sume-managed products such as Avatar, audio, and media tools have fixed published rates. The Image and Video Routers pass through the provider's list rate times Sume's margin; the Video Router docs state that margin as list × 1.25 on every video model. Per-model rates are in the catalog (GET /v1/catalog) and on API pricing.
| Product | Rate |
|---|---|
| Create avatar | $0.95 per avatar |
| Avatar Video, standard, no product image | $0.184 per second |
| Avatar Video, standard, with product image | $0.194 per second |
| Video Router models | Provider list rate × 1.25 |
How do I cap what one run can spend?
Unattended runs carry their own ceiling. A Format run takes generation_spend_cap_usd (up to the platform maximum of $500) and can never spend past its effective cap; an Agent Completion requires the field. Every receipt reports the cap and what the run actually spent (usage.billable_amount_usd_micros). See Create a run for the rules.
Which plan do I need for the API and MCP?
API and MCP access starts with Pro and is included in every higher plan. Free covers image generation and limited Sume Agent access. Startup adds a shared team workspace, and Scale adds the highest self-serve concurrency and priority support. Enterprise is for custom terms and invoice or PO billing.