Anthropic announced Claude 3.7 Sonnet on February 24, 2025, with an extended thinking mode that lets the model spend additional effort on a response. “As long as you want” was not a promise of unlimited computation: developers could set a thinking budget, and users could toggle the mode. Current controls and model support depend on the Claude model and API generation.
What Anthropic announced
Claude 3.7 Sonnet combined ordinary responses with an optional extended thinking mode. Anthropic said the mode did not switch to a different model or a separate strategy. In its February 24, 2025 announcement, the company put it this way: “Extended thinking mode isn’t an option that switches to a different model with a separate strategy. Instead, it’s allowing the very same model to give itself more time, and expend more effort, in coming to an answer.” Anthropic’s announcement
![Claude AI for Beginners Bible: [5 in 1] The Ultimate Guide to Automate Your Work, Save Hours Every Week, and Use AI for Real-World Results](https://m.media-amazon.com/images/I/415+fSJacsL._SL160_.jpg)
At launch, users could toggle extended thinking on or off, while developers could configure a thinking budget. Anthropic described the visible thought process as a research preview. The launch announcement said Claude 3.7 Sonnet was available through Claude.ai and the API, and named Pro, Team, Enterprise, and API users as eligible; those are launch-era details, not confirmation of current plan terms. Anthropic’s announcement
What “as long as you want” means in practice
The phrase describes adjustable effort, not unlimited runtime. In the current API documentation, manual extended thinking is configured with thinking: {type: "enabled", budget_tokens: N}. The budget is a target of at least 1,024 tokens and, except in the documented interleaved-thinking case, must be smaller than max_tokens. Anthropic’s extended thinking documentation
Recommended Free Tools
Thinking tokens count as output tokens and share the max_tokens limit with the final response. That means allocating more tokens to thinking can affect the room available for the answer and can increase usage and response time. A thinking budget is a control within those limits, not a guarantee of a particular duration or improved result. Anthropic’s extended thinking documentation
Manual extended thinking and adaptive thinking
Anthropic’s current API documentation distinguishes manual extended thinking from adaptive thinking. Which approach applies depends on the model generation, so check support for the exact model you plan to call before using an API parameter.
| Mode | How it works | Configuration and availability | Token and response considerations |
|---|---|---|---|
| Manual extended thinking | You enable thinking and provide a target budget. | Uses thinking: {type: "enabled", budget_tokens: N}. It remains relevant for models that support only this mode. Anthropic says it is deprecated on Claude 4.6 models and rejected by Claude 4.7 and later models. |
Thinking tokens count toward max_tokens alongside the final response. A thinking summary may be returned, but the displayed text is not raw chain of thought. |
| Adaptive thinking | The model determines whether and how deeply to think based on configuration and task complexity. | Use it where supported; newer model generations may require it instead of manual extended thinking. | Thinking still uses output tokens. The amount of thinking and the response’s timing can vary with the task and configuration. |
These distinctions and model-generation rules are described in Anthropic’s extended thinking documentation and its adaptive thinking documentation. Because model support and parameter rules can change, consult the current documentation for the target model before implementing either mode.
What users see—and what the feature is for
A thinking display should not be treated as a verbatim record of a model’s private reasoning. Anthropic says the text shown may be summarized or omitted, and is not raw chain of thought. Anthropic’s extended thinking documentation
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallAnthropic identifies math, coding, analysis, and long-running agentic tasks as possible uses for extended thinking. These are vendor-stated use cases, not a guarantee that enabling it will improve every answer. For a simple request, the extra effort may be unnecessary; for a complex task, it may give the model more room to work through the problem. Anthropic’s extended thinking documentation
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Bottom line
Claude 3.7 Sonnet’s 2025 launch introduced a configurable way to give the same model more room to work on a response. The setting was never literally unlimited, and Anthropic’s current API distinguishes the older manual budget from adaptive thinking, with support varying by model generation.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
