Do subagents count toward the Claude Code usage limit?
Updated
Yes. The Claude Code docs say the /usage plan breakdown shows 'what counts against your plan limits', with usage 'attributed to skills, subagents, plugins, and individual MCP servers'. Each subagent runs in its own context window, so its tokens come on top of the main conversation, and 'Agent teams use approximately 7x more tokens than standard sessions when teammates run in plan mode.' To spend less, set model: haiku for simple subagent tasks.
What the official docs say
/usage also shows a breakdown of what counts against your plan limits: Attribution: recent usage attributed to skills, subagents, plugins, and individual MCP servers
Quoted from Claude Code docs: Manage costs effectively Each subagent runs in its own context window with a custom system prompt, specific tool access, and independent permissions
Quoted from Claude Code docs: Create custom subagents Token usage scales with the number of active teammates and how long each one runs
Quoted from Claude Code docs: Manage costs effectively Agent teams use approximately 7x more tokens than standard sessions when teammates run in plan mode.
Quoted from Claude Code docs: Manage costs effectively For simple subagent tasks, specify model: haiku in your subagent configuration.
Quoted from Claude Code docs: Manage costs effectively
How it is counted
Subagents are metered against your plan. The cost guide says that on a Pro, Max, Team or Enterprise plan '/usage also shows a breakdown of what counts against your plan limits: Attribution: recent usage attributed to skills, subagents, plugins, and individual MCP servers'.
Each one is a separate context. The subagent docs say 'Each subagent runs in its own context window with a custom system prompt, specific tool access, and independent permissions', so a subagent's tokens are added to, not shared with, the main conversation. For agent teams, 'Token usage scales with the number of active teammates and how long each one runs', and 'Agent teams use approximately 7x more tokens than standard sessions when teammates run in plan mode.'
They can still save main-context tokens. Verbose tool output stays in the subagent's context while only a summary returns to your main conversation, which keeps later turns cheaper. The cost guide's lever for the subagent side: 'For simple subagent tasks, specify model: haiku in your subagent configuration.'
| Feature | Counted toward the session / weekly limit? | Source |
|---|---|---|
| Subagents | Yes. Listed in the /usage attribution breakdown of what counts against plan limits | [1] |
| Subagent context | Own context window per subagent, added on top of the main conversation | [2] |
| Agent teams | Yes. Approximately 7x more tokens than standard sessions when teammates run in plan mode | [1] |
| Skills, plugins, MCP servers | Yes. Attributed in the same breakdown | [1] |
| Cheaper subagents | Set model: haiku for simple subagent tasks | [1] |
FAQ
Do subagents save usage or cost more?
Both can be true. Their tokens count against your plan, and each runs in its own context window. But verbose output stays in the subagent's context while only a summary returns to your main conversation, so the main thread stays smaller for later turns.
How much more do agent teams use?
The cost guide says 'Agent teams use approximately 7x more tokens than standard sessions when teammates run in plan mode', and that 'Token usage scales with the number of active teammates and how long each one runs'.
Where do I see what my subagents used?
Run /usage in Claude Code. On a Pro, Max, Team or Enterprise plan it shows 'a breakdown of what counts against your plan limits', including recent usage attributed to skills, subagents, plugins and individual MCP servers.
How do I make subagents cheaper?
The cost guide says: 'For simple subagent tasks, specify model: haiku in your subagent configuration.' Keep heavier models for the subagents that need them.
Sources
- Claude Code docs: Manage costs effectivelyverified 2026-09-17
- Claude Code docs: Create custom subagentsverified 2026-09-17
Want the next Claude reset on your phone?