# Instance types & specifications

Mithril offers several instances of various NVIDIA GPU types fit for your workloads. We will continue to add instance types over time.

Depending on your instance type, Ethernet or InfiniBand interconnect may be available.

## VM image

You can choose between one of the supported VM images. We recommend `Ubuntu24.04 2025.10.10` for the latest software, best performance, and updated security. This image is the default if none is specified in the launch specification.

<table><thead><tr><th width="225.6015625">Image name</th><th width="143.71484375">NVIDIA driver</th><th width="99.4375">CUDA</th><th width="266.5625">Notes</th><th width="166.41015625">NVIDIA container toolkit</th><th width="161.65625">OS</th><th data-hidden>Supported regions<select multiple><option value="FlBcdazsdepl" label="us-central1b" color="blue"></option><option value="HATYHkH84Ds7" label="us-central1-b" color="blue"></option><option value="eT95vvpRfwuX" label="us-central2-a" color="blue"></option><option value="RwntkCb0H5Jj" label="eu-central1-b" color="blue"></option><option value="1i2W3ZxlXaHm" label="all" color="blue"></option></select></th></tr></thead><tbody><tr><td>Ubuntu24.04 2025.10.10</td><td>580.95.05</td><td>13.0</td><td>Improved fileshare performance</td><td>1.17.8</td><td>Ubuntu 24.04 LTS</td><td></td></tr><tr><td>Ubuntu24.04 2025.08.01</td><td>570.86.10</td><td>12.8</td><td>Does not support executing binaries from persistent storage</td><td>1.17.8</td><td>Ubuntu 24.04 LTS</td><td><span data-option="1i2W3ZxlXaHm">all</span></td></tr><tr><td>Ubuntu22.04 2024.07.24</td><td>535.183.01</td><td>12.2</td><td></td><td></td><td>Ubuntu 22.04</td><td><span data-option="1i2W3ZxlXaHm">all</span></td></tr></tbody></table>

You can use [startup scripts](/compute-and-storage/startup-scripts.md) to customize your environment during the startup process of your instances.

## Hardware specifications

Each instance comes with its own ephemeral storage (outlined below) with NVMe SSD. This storage will exist until the instance has been terminated or pre-empted. There is no cost to use the ephemeral storage. Ephemeral storage is automatically mounted for you (and re-mounted after preemptions) at `/mnt/local`.

{% hint style="info" %}
Achieve 25% faster start-up times when configuring your instances with RAM lower than the default! RAM can be specified in the console or API, with smaller amounts enabling faster start-up time. Note the compute price does not change based on how much RAM you have — pricing is per GPU hour.
{% endhint %}

<table data-full-width="true"><thead><tr><th width="114.75">Instance</th><th width="109.63671875">GPU type</th><th width="85.25390625">GPUs</th><th width="115.09765625">VRAM per GPU (GB)</th><th width="94.77734375">CPU cores</th><th width="111.80078125">System RAM (GB)</th><th width="108.12109375">Boot disk (GB)</th><th width="150.54296875">Ephemeral NVMe (GB)</th></tr></thead><tbody><tr><td>1x A100 80GB SXM</td><td>A100 SXM</td><td>1</td><td>80</td><td>26</td><td>190</td><td>256</td><td>1,750</td></tr><tr><td>2x A100 80GB SXM</td><td>A100 SXM</td><td>2</td><td>80</td><td>52</td><td>380</td><td>256</td><td>3,500</td></tr><tr><td>4x A100 80GB SXM</td><td>A100 SXM</td><td>4</td><td>80</td><td>102</td><td>760</td><td>256</td><td>7,000</td></tr><tr><td>8x A100 80GB SXM</td><td>A100 SXM</td><td>8</td><td>80</td><td>204</td><td>1,520</td><td>256</td><td>14,000</td></tr><tr><td>8x H100 80GB SXM</td><td>H100 SXM</td><td>8</td><td>80</td><td>168</td><td>1,814</td><td>256</td><td>12,500</td></tr><tr><td>8x H200 141GB SXM</td><td>H200 SXM</td><td>8</td><td>141</td><td>168</td><td>1,600</td><td>256</td><td>1,120</td></tr><tr><td>8x B200 192GB SXM</td><td>B200 SXM</td><td>8</td><td>192</td><td>232</td><td>3,584</td><td>256</td><td>N/A</td></tr></tbody></table>

#### Memory (RAM) configuration

To improve Spot startup times by up to 25%, we recommend launching bids with less RAM configured in the launch specification. This is ideal for light workloads or development use cases that require fast startup but low resource utilization.&#x20;

Use the console or API to launch instances with less RAM.

At any point after launch, RAM can be adjusted from 100GB up to the instance type's default. In order to adjust the RAM for an order, the bid must be paused and resumed, which will restart all instances in the order.

By default, instances are launched with the maximum memory (RAM) allowed for the instance type.&#x20;

<figure><img src="/files/iJazLgMqKBYKehPNMiWu" alt=""><figcaption></figcaption></figure>

Note that this feature is only available for Spot instances at this time.


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.mithril.ai/compute-and-storage/instance-types-and-specifications.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.