Anthropic Says Drop "Think Hard" From Your Opus 5.5 Prompts
Opus 5.5 is now the default model on paid Claude plans. Anthropic advises starting at a medium effort level and not asking the model to think harder.
Anthropic, the company behind the Claude chatbot, has made Claude Opus 5.5 the default model on its paid Pro, Max, Ultra and Team plans. Alongside the switch, it published a guide on how to write requests for the new model. Its main advice is to stop typing phrases like "think hard" and "think carefully," and to begin at the Medium effort level. Effort is a setting that controls how long the model reasons before it answers.
The reason is that Opus 5.5 always thinks before it answers and decides for itself how long to spend. In the company's tests, removing those phrases made replies start sooner, with no clear drop in quality. Anthropic says Opus 5.5 at Medium matched or beat the older Opus 5 at High. Higher settings think longer and cost more, but they do not always produce better work. The company suggests moving to High only when Medium stalls.
For regular users, the change is simple. Old habits of urging the model to try harder now mostly add waiting time. Anthropic also notes that pushing the cheaper Sonnet 5.5 to its high settings brings the cost per task close to Opus 5.5. That removes much of the reason to pick Sonnet for hard tasks. Every figure here comes from Anthropic's own testing. Outside tests have not yet confirmed them.
Source: claude.com
Free course
Stop reading about AI — start building with it
The free Claude Code course: your first site, tool or game — no coding. No upsells, no cross-sells — nothing to buy here.
Start free →▌ Related guides

Author
Evgenii Arsentev
PhD · Founder & CEO, Dad4Kids
Articles · Latest articles