◈AI Codex
Foundation Models & LLMs◆ ClaudeOperatorsDevelopers

Extended Thinking

Also: extended thinking mode, Claude thinking mode

Extended thinking was the setting that let Claude reason step by step before answering, with a fixed thinking budget you chose. On current models it has been replaced by adaptive thinking: the model decides how much to think, and you control depth with an effort level (low to max) instead of a token budget. The idea is unchanged, so more effort still means better answers on hard problems at higher cost, but the old budget setting now returns an error on the Claude 5.x models.

◎

In practice

If you have code calling `thinking: {type: "enabled", budget_tokens: 10000}` against `claude-opus-5`, it will fail. Replace it with `output_config: {effort: "xhigh"}` for demanding agentic and coding work, or drop the parameter entirely to get the `high` default. To see the reasoning, which is hidden by default on current models, pass `thinking: {"type": "adaptive", "display": "summarized"}`.

Related concepts

Where Extended Thinking shows up

1 article