> For the complete documentation index, see [llms.txt](https://help.swarmia.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.swarmia.com/features/ai-tools/ai-cost.md).
# AI cost (beta)
Available at [AI tools → AI cost](https://app.swarmia.com/ai/cost/overview)
AI cost shows how much your organization's AI coding tool usage is worth and breaks it down by team, by person, and by the work your teams deliver. It's based on [token value](#token-value): the list price of the tokens your tools consume.
## Token value
Token value is the list price of the tokens used, including the harness (overhead the agent adds, like system prompts and tool calls), regardless of what you actually pay. On subscription plans where heavy usage doesn't increase your bill, this tells you what that usage is worth. All amounts are shown in US dollars.
Where the data comes from depends on the tool:
* **GitHub Copilot:** [GitHub AI credit](https://docs.github.com/en/copilot/concepts/billing/usage-based-billing-for-individuals) usage from the [GitHub Copilot integration](/settings/integrations/ai-coding-tool-integrations/github-copilot-integration.md). Data is available from June 19, 2026 onwards.
* **Cursor:** Usage data from the [Cursor integration](/settings/integrations/ai-coding-tool-integrations/cursor-integration.md). Includes the list price of token usage and the Cursor fee, without discounts.
* **Claude Code:** Anthropic's "estimated cost" from the [Claude Code integration](/settings/integrations/ai-coding-tool-integrations/claude-code-integration.md) (supports both the [OpenTelemetry monitoring](/settings/integrations/ai-coding-tool-integrations/claude-code-integration.md#opentelemetry-monitoring) and the [Analytics API](/settings/integrations/ai-coding-tool-integrations/claude-code-integration.md#analytics-api)). Includes the list price of token usage, as well as harness costs such as prompt caching and tool use.
Token value isn't your invoice. It's a single comparable number that lets you aggregate AI usage across tools, teams, people, and work items. See [why we report token value instead of token counts](#why-do-you-report-the-token-value-instead-of-the-token-count) for the reasoning.
## The AI cost page
Pick a team and a timeframe, and the page shows:
* **Token value chart:** Claude Code, Cursor, and GitHub Copilot token value over time, with the total for the timeframe compared to the previous one.
* **Teams tab:** Token value per tool for each team, alongside the number of [AI-assisted pull requests](/features/ai-tools/ai-tool-detection-and-filters.md) merged and the number of people who used AI tools. Hover over the **Users** column to see each person's token value, and select a team to drill into its child teams.
* **Users tab:** The same breakdown per person for the selected team.
* **Work tab:** An embedded [focus summary](/features/focus/focus-summary.md) scoped to the selected team and timeframe, so you can see which issues and initiatives the spend went to.
Both the Teams and Users tabs can be downloaded as a CSV.
The breakdowns are based on your current team members. Historical team memberships aren't supported: if people switch teams, their past token value moves with them to their current team. People who don't belong to any team are left out of the Users tab.
## AI spend in the focus summary
The [Work tab](#the-ai-cost-page) of the AI cost page and the [focus summary](/features/focus/focus-summary.md) include an **AI spend** column next to each issue or initiative. It shows the [token value](#token-value) from GitHub Copilot, Cursor, and Claude Code attributed to that work and its share of the total in the view. You can hover on the values to see a breakdown by AI tool and the time range of the data.
To attribute token value to work, Swarmia spreads each person's daily AI usage across the issues and pull requests they contributed to that day, in proportion to their [effort](/definitions/developer-effort-ftes.md). If someone uses AI tools on a day without attributable work, that usage is left out of the focus summary breakdown.
## AI spend in the issue pop-up
You can open any issue in Swarmia to see a spend breakdown by AI tool and contributor.
## Setup
AI cost is built on the [AI coding tool integrations](/settings/integrations/ai-coding-tool-integrations.md), so no separate setup is needed:
* **GitHub Copilot** requires the [GitHub Copilot integration](/settings/integrations/ai-coding-tool-integrations/github-copilot-integration.md).
* **Cursor** requires the [Cursor integration](/settings/integrations/ai-coding-tool-integrations/cursor-integration.md).
* **Claude Code** requires the OpenTelemetry setup described in the [Claude Code integration](/settings/integrations/ai-coding-tool-integrations/claude-code-integration.md#opentelemetry-monitoring).
For token value to land on the right person, each contributor's AI tool identities need to be connected in the [Contributor settings](/settings/organization/contributors.md).
## Frequently asked questions
### Why don't I see costs for today?
Like other AI tool data in Swarmia, cost data is synced periodically, so the numbers can be up to a day behind.
### Why is someone's cost attributed to the wrong person or missing?
Check that their AI tool identities are connected to the right contributor in the [Contributor settings](/settings/organization/contributors.md). Swarmia attributes cost to the merged contributor, so unmerged identities can leave cost unattributed.
### Why doesn't the token value match our AI tool invoice?
Token value is the list price of the tokens and harness (overhead the agent adds, like system prompts and tool calls) used, not your billed cost. On subscription plans like Claude Team, token usage doesn't map directly to what you pay. Seat license costs aren't included either.
### Why do you report the token value instead of the token count?
Adding up tokens into a single number would lead to an apples-to-oranges comparison:
1. Different LLMs price tokens differently. Claude Fable tokens are [over three times as expensive](https://platform.claude.com/docs/en/about-claude/pricing) as Claude Sonnet tokens, which arguably also reflects the value you get from them.
2. Each model uses its own tokenizer, so the same work can produce different token counts. For example, [Claude Opus 4.7 produces about 30% more tokens](https://www.anthropic.com/news/claude-opus-4-7) than Claude Opus 4.6 for the same exact text.
3. Different token types (input, output, cache read, and cache write) are priced differently — an output token can be [50 times the cost](https://platform.claude.com/docs/en/about-claude/pricing) of a cache-read token.
Token value converts usage into one comparable metric: the list price of the tokens used. This allows you to compare and aggregate AI spend across teams, people, issues, pull requests, and AI tools.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## 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, and the optional `goal` query parameter:
```
GET https://help.swarmia.com/features/ai-tools/ai-cost.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
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.