Adaptive Thinking
Adaptive thinking is how Claude decides for itself how much reasoning to apply to a given request. Simple questions get fast, direct answers; complex or ambiguous ones trigger more deliberate reasoning first. It is the default on Claude Opus 5, Claude Sonnet 5 and Claude Fable 5, and available but off by default on Claude Opus 4.8, 4.7, 4.6 and Sonnet 4.6 until you set `thinking: {type: "adaptive"}`. It replaced extended thinking, where you enabled reasoning manually and set a token budget.
In practice
You ask Claude a quick factual question and it answers in one sentence. You ask it to help you think through a tricky hiring decision and it slows down, considers multiple angles, and works through its reasoning first. You didn't change anything — Claude calibrated the depth on its own. When you want to influence that calibration, the control is `effort` (low, medium, high, xhigh, max), not a thinking toggle.
Related concepts
Where Adaptive Thinking shows up
3 articlesClaude doesn't apply the same effort to every question. Here's what adaptive thinking is, how it works, and why it matters for the outputs you get.
Claude Opus 5.5 shipped on September 22, 2026. It costs $4/$20 per million tokens, cache reads drop 60% to $0.20, and Anthropic says it performs at the level of Fable 5.1 on most work. It also defaults to medium effort, cannot turn thinking off, and rejects forced tool use. Here is who should switch, what breaks, and how to test it on your own tasks.
Claude Sonnet 5.5 shipped on September 28, 2026 at $2/$10 per million tokens, with output more than 30% faster than Sonnet 5 and up to 30% lower cost per task. It also changes how thinking is turned off, rejects forced tool use, and retires Sonnet 4.5 on November 30. Here is the checklist.