Intermediate8 min
Effort Levels — The Reasoning Depth Slider
Effort levels control how many thinking tokens Claude allocates before responding. From the model selection angle, the right effort level is as important as the right model — and the two choices compound.
Quick Reference
- →5 levels: low, medium, high (default), xhigh (Opus 4.7 default), max (one session only)
- →Effort controls thinking token budget — more effort = deeper reasoning = more tokens
- →Set: /effort, --effort flag, effortLevel in settings.json, CLAUDE_CODE_EFFORT_LEVEL env var
- →ultrathink in prompt: overrides effort level upward for one specific response
- →Effort is session-wide — ultrathink is per-prompt
- →High effort teams: cost of effort choices compounds significantly across many sessions
What Effort Actually Controls
The effort level sets the thinking token budget — how many tokens Claude allocates to internal reasoning before producing a response. Higher effort means Claude can explore more approaches, reason through more edge cases, and produce a more considered answer. It also means more tokens consumed per response.
| Level | Thinking depth | Default for | Token cost |
|---|---|---|---|
| low | Minimal — quick answers | Simple lookups | Lowest |
| medium | Balanced thinking for clear tasks | — | Low |
| high | Deep reasoning, thorough analysis | Most models | Standard |
| xhigh | Extra deep, broader exploration | Opus 4.7 | High |
| max | Maximum reasoning, full budget | One session only | Highest |