Who is eligible
- Individual users only. The trial is not available to companies. When you sign up and set up billing, you must choose to pay as an individual user.
- First-time trial users only. The trial can be activated only once per user. If you add a payment card that was previously used to activate a trial on any account, the trial will not be offered and you will need to use a different card.
What you get
- 30 days of access to the Nebius AI Cloud platform.
- $5 in credit to spend on the resources of your choice.
Available regions
The trial is available only in theeu-north1 region.
Quotas
Certain Nebius AI Cloud resources have specific quotas for free trial users.| Quota | Value,eu-north1 |
| Compute | |
| Total number of disks | 1 |
| Total SSD storage capacity | 1280 GiB |
| Preemptible virtual machines | 1 |
| Object Storage | |
| Number of buckets | 1 |
| Total size of objects in Standard storage | 100 GiB |
| Virtual Networks | |
| Total number of occupied public IPv4 addresses | 1 |
You can check all other quotas on the Administration → Limits → Quotas page of the web console.
When the trial ends
At any time throughout the trial, you can decide how to proceed when the trial ends (your credit runs out or 30 days have passed):- Stop all usage. Deactivate any active resources. No further charges apply.
- Continue to paid usage. Keep your resources running. The payment card you added for the trial will be charged $25 for the initial top-up, and your existing resources continue to run under the pay-as-you-go billing model. All quotas will be updated to their default values.
Make sure to stop or delete any resources you don’t need before switching to paid usage to avoid being charged for idle capacity.
Example use cases
You can use the free trial to run small, low-friction experiments that fit within the $5 credit and 30-day window.Run inference
Use a Serverless AI endpoint to run inference for a model that is not a large language model (LLM), for example, video or image generation, or bring your own model.Fine-tune a model
Use a Serverless AI job and Axolotl to fine-tune the Qwen/Qwen2.5-0.5B large language model.Test an idea
Use the JupyterLab® application to quickly test an idea that requires heavy GPU computation without the need to set up infrastructure.“Jupyter” and the Jupyter logos are trademarks or registered trademarks of LF Charities, used by Nebius B.V. with permission.