# AI providers and models

> **Note:**
> Viewing the page requires the [View AI Analytics](roles.html#permissions_view_ai_analytics) or [Manage organization](roles.html#permissions_manage_org) permission.
>
> * The following default roles include the [View AI Analytics](roles.html#permissions_view_ai_analytics) permission: [Org viewer](roles.html#org_viewer), [AI admin](roles.html#ai_admin), [AI viewer](roles.html#ai_viewer), [AI analytics viewer](roles.html#ai_analytics_viewer).
>
> * The only default role that includes the [Manage organization](roles.html#permissions_manage_org) permission is [Org admin](roles.html#org_admin).

The AI providers and models segment lets you track token usage in your organization, by AI provider and model. This data can help you understand token spending, compare usage patterns across providers and models, and make informed operational and budgeting decisions.

> **Note:**
> Token usage data is currently available only for JetBrains AI. Once centrally managed access to third-party AI providers becomes available at the organization level, their usage will also appear in these charts and tables.

## Using this data

The AI providers and models segment helps you:

* Identify which providers and models are most heavily used in your organization.

* Optimize your AI strategy by shifting workloads to more cost-effective providers or models.

* Plan budgets based on actual usage data.

## Data sources

Token usage is tracked from the use of AI tools through a JetBrains [license with AI](ai-management.html#license-with-ai). This usage is attributed to JetBrains AI as the provider, regardless of which AI tool is used. For example, the use of Codex in AI Assistant through a JetBrains license with AI counts towards JetBrains AI usage.

> **Note:**
> The reported token consumption includes all token types: input, output, thinking, cache read, and cache write.

The following usage is excluded from the token usage charts and tables:

* User-configured [third-party AI providers](https://www.jetbrains.com.cn/en-us/help/ai-assistant/use-custom-models.html#provide-your-own-api-key). With this approach, each user can set their own access to third-party providers through API keys or subscriptions directly in AI tools. This type of access is not managed centrally by your organization, so it is not included in the token usage data.

* In-editor features. The use of features such as [Next edit suggestions](https://www.jetbrains.com.cn/en-us/help/ai-assistant/next-edit-suggestions.html) and [Code Completion](https://www.jetbrains.com.cn/en-us/help/ai-assistant/code-completion.html) is not included, as they don't consume paid tokens.

## Data filtering

You can filter the data by the following criteria:

* In the time range filter, select one of the predefined filters that show the last specified period, or select a custom date range. The data includes both the first and the last day of the selected period.

* In the provider filter, which displays All providers by default, select the AI providers you want to show the data for. When no provider is selected, the data is shown for all providers.

* In the user filter, which displays All users by default, search for and select [users](users.html) or [groups](groups.html). Selecting a group includes all of its members and members of its subgroups. Expand a group to select its subgroups.

These filters apply to both the [Token usage by provider](#token-usage-by-provider) chart and the [Token usage by model](#token-usage-by-model) table.

Filters of different types are combined with a logical `AND`. For example, data will match both the selected date range and the user the data is attributed to.

Multiple selections of the same filter are combined with a logical `OR`. For example, selecting a group along with an individual user shows data for that user and all members of the group.

Filter selections are preserved across analytics segments. The filters that you set in one segment are automatically preselected when you switch to another segment.

## Token usage by provider

This chart displays token consumption per AI provider for your selected time period.

![Token usage by provider chart](images/token-usage-by-provider.png)

In the legend, the indicator next to the token count shows the token consumption trend compared to the previous period of the same length. For example, if you select a 30-day period, it is compared to the 30 days prior to the start of your selected period.

The token consumption trend, if detected, can show one of the following indicators:

* ![Token consumption increase icon](images/token-consumption-increase.svg) Increase

* ![Token consumption decrease icon](images/token-consumption-decrease.svg) Decrease

* ![Equal token consumption icon](images/token-consumption-equal.svg) No change in consumption or not enough data for comparison > **Tip:** > When you hover over the ![Equal token consumption icon](images/token-consumption-equal.svg) icon, the tooltip shows whether the consumption is equal or there isn't enough data for comparison.

## Token usage by model

The table shows a breakdown of token usage by AI model for your selected time period. Currently, the data includes token consumption for models available through JetBrains AI only.

![Token usage by model table](images/token-usage-by-model.png)