Saturday, November 14, 2026 Edition no. 30 · Evening

Artificial intelligence news from Brazil, in English

ProductsMedium importance

Anthropic recommends dropping “think hard” from Claude Opus 5.5 prompts, now the default model on paid plans, and starting at medium effort

In a guide published on the Claude website, Anthropic says Opus 5.5 has become the default model on the Pro, Max, Ultra and Team plans and recommends it as the everyday model for research, analysis and writing. According to The Decoder, the company asks users to cut phrases such as “think hard” and “think carefully”: the model already adjusts on its own how much it reasons, and in internal tests responses without those phrases started faster, with no measured loss in quality. Anthropic also recommends starting at the “Medium” effort level, the default, which in internal tests matched Opus 5 on “High”. For those who used Sonnet day to day and saved Opus for heavy tasks, the company says it is worth reconsidering, because a more capable model can finish the task with less work and offset much of its higher price per token.

Joãozinho AI Published in edition no. 30 of Noticiário do Joãozinho ·

Translated by Joãozinho AI from the original Brazilian Portuguese edition. Facts, numbers and sources are the same as in the original; quotes, video titles and source names stay in their original language. Read the original in Portuguese

Original source

Anthropic (Claude)The Decoder: the habits Anthropic says to drop

Use the ← → arrow keys to change stories