feature
Effort & fast mode
Two dials — how hard Claude thinks about a task, and how fast it works — so you can match the model to the job.
Not every task needs the same care. **Effort** sets how much thinking Claude does before it answers — dial it up for a thorny bug, down for a quick edit. **Fast mode** is the other dial: it trades a little money for noticeably more speed when you want answers sooner. On Claude Opus 4.8 the default is already high effort, so the hardest work gets deep thinking without you asking.
Two separate dials control how Claude works, and it’s worth knowing which is
which. Effort is about depth of thought: at higher effort Claude reasons
longer before it acts, which is exactly what you want for a subtle bug or a design
with real trade-offs. Claude Opus 4.8 runs at high effort by default, and
/effort xhigh pushes it further for the hardest tasks. Lower effort is fine — and
faster — for small, obvious edits.
Fast mode is a different trade entirely: it doesn’t change how carefully Claude
thinks, it changes how quickly it responds. Toggle it with /fast to get roughly
2.5× the speed for a higher per-token rate — handy when you’re iterating in tight
loops and waiting is the bottleneck. Turn it back off when you’re done and want the
standard rate again.
The rule of thumb: use effort to match the difficulty of the task, and fast mode to match your impatience. They stack — you can run a hard task at high effort and in fast mode if both depth and speed matter.
why it helps Spend the model's effort where it pays off — and get speed when you're in a hurry.
examples
/effort xhigh/fast/model claude-opus-4-8tips & best practices
- Reach for
/effort xhighon the genuinely hard stuff — a subtle bug, a tricky design — and leave it on high for everyday work. - Turn on
/fastwhen you're iterating quickly and want answers sooner; it costs a bit more per token for the extra speed. - Higher effort means Claude thinks longer before acting — great for correctness, slower for trivial edits.