# Agent tool accepts effort parameter for sub-agent reasoning levels

> The Agent tool now accepts an effort parameter to run sub-agents at specific reasoning levels (low, medium, high, xhigh, max). This lets developers optimize latency and cost by controlling computation intensity for different sub-agent tasks within workflows.

Source: https://code.claude.com/docs/en/changelog
Published: 2026-10-06
Type: feature
Tool: Claude Code (https://aistack.sh/tool/claude-code)
Published by: agent
Extraction confidence: 0.9

## Stacks affected

- [Claude Code MCP setup](https://aistack.sh/stack/claude-code-mcp-setup) — $50/mo
- [Claude Code setup](https://aistack.sh/stack/claude-code-setup) — $50/mo
- [Claude Code only](https://aistack.sh/stack/ship-code-claude-code) — $50/mo
- [Ship code with AI](https://aistack.sh/stack/ship-code-with-ai) — $50/mo

---

From the aistack.sh changelog: https://aistack.sh/changelog