> See the full index of documentation pages at https://docs.zettabyte.com/llms.txt.

# Choose a product

> Find the Zettabyte product that fits the task: GPU instances for running work, or one API for calling models.

Source: https://docs.zettabyte.com/docs/get-started/choose-a-product

Last updated: 2026-11-05

To rent GPUs and run models or other work on them, choose zCLOUD. To call models through one API, choose zSTUDIO. Both products are reached from the same console and account, as the **AI Infrastructure Rental** and **MaaS via API** workspaces.

## GPU cloud (zCLOUD)

zCLOUD is a GPU cloud for developing, training, fine-tuning and serving AI models, on GPU instances that customers rent. It is billed by instance and GPU hours. For a developer or team that wants to run its own GPU work. Start with [What zCLOUD is](https://docs.zettabyte.com/docs/zcloud/introduction/what-zcloud-is).

## Model as a Service (zSTUDIO)

zSTUDIO gives access to a choice of models through one API, using the OpenAI SDK. Calling models is billed by token usage and needs no instances to rent. For a developer or team building an app or agent that calls models. Start with [What is Model as a Service (MaaS)?](https://docs.zettabyte.com/docs/zstudio/introduction/what-is-maas).

## Next steps

- [Sign in and accounts](https://docs.zettabyte.com/docs/get-started/sign-in-and-accounts): when ready to sign in or add a colleague to the account.
- [API keys](https://docs.zettabyte.com/docs/accounts-and-roles/api-keys): when an app or script needs to authenticate.
- [Accounts and roles](https://docs.zettabyte.com/docs/accounts-and-roles): when deciding which console controls a member sees.
- [Billing](https://docs.zettabyte.com/docs/billing): when checking the wallet balance and charges.
