Anthropic’s Claude AI Launches New Model Capable of Extended Thinking
The New Model Doesn’t Come Free
Anthropic’s Claude AI has launched a new model, known as Claude 3.7 Sonnet, which is capable of extended thinking and deeper responses. The new model uses advanced reasoning and greater processor time to evaluate a question in a step-by-step process and produce a detailed result.
Accessing the Extended Thinking Mode
The extended thinking mode is accessible through the Claude website and the API for developers. However, it doesn’t come free. Users must have a Pro or Team subscription to tap into the extended thinking option.
The Benefits of Extended Thinking
The extended mode is particularly adept at tackling difficult math and coding problems, as well as front-end web development. In this mode, Claude "self-reflects" before providing an answer, taking the time to develop its response and better handle tasks that involve math, physics, instruction-following, and coding. Along the way, Claude shows users the steps it took to arrive at its solution.
Early Testing and Evaluation
In its announcement, Anthropic touted the performance of Claude 3.7 Sonnet in early testing and use. One site found the new model significantly improved at handling complex codebases and using advanced tools. Another successfully used Claude to build sophisticated web apps and dashboards from scratch. In a third evaluation, Claude created production-ready code with quality designs and dramatically fewer errors.
Comparison with OpenAI’s o1
Testing of agentic tools put Claude 3.7 Sonnet ahead of the 3.5/3.6 version of Sonnet and OpenAI’s o1 model. Testing in software engineering placed Claude 3.7 Sonnet at the top among the previous version, OpenAI’s o1, and DeepSeek R1.
How to Try Claude 3.7 Sonnet and Claude Code
To try Claude’s new extended mode, users must be a Pro or Team subscriber. Head to the website, click the drop-down menu for the models, and select Claude 3.7 Sonnet. Change the thinking mode from Normal to Extended, enter and submit your request.
Claude 3.7 Sonnet in Action
In response to a request, Claude displayed each line of HTML and CSS code as it was being generated. When Claude finished its work, users can view the page in HTML mode or in preview mode.
Claude Code for Developers
Claude 3.7 Sonnet also offers an innovation for developers. With this new version, Anthropic has introduced a command-line tool for agentic coding, known as Claude Code. This allows developers to assign intricate engineering tasks to Claude directly from the terminal. Available for now as a limited research preview, Claude Code can search for and read code, edit files, write and run tests, and commit and push code to GitHub.
Conclusion
Anthropic’s Claude AI has launched a new model capable of extended thinking and deeper responses. The new model uses advanced reasoning and greater processor time to evaluate a question in a step-by-step process and produce a detailed result. While it doesn’t come free, users with a Pro or Team subscription can access the extended thinking option.
Frequently Asked Questions
Q: What is the new model capable of?
A: Claude 3.7 Sonnet is capable of extended thinking and deeper responses.
Q: How do I access the extended thinking mode?
A: The extended thinking mode is accessible through the Claude website and the API for developers. Users must have a Pro or Team subscription to tap into the extended thinking option.
Q: What are the benefits of extended thinking?
A: The extended mode is particularly adept at tackling difficult math and coding problems, as well as front-end web development.
Q: How does Claude 3.7 Sonnet compare to OpenAI’s o1?
A: Testing of agentic tools put Claude 3.7 Sonnet ahead of the 3.5/3.6 version of Sonnet and OpenAI’s o1 model.
Q: How can I try Claude 3.7 Sonnet and Claude Code?
A: To try Claude’s new extended mode, users must be a Pro or Team subscriber. Head to the website, click the drop-down menu for the models, and select Claude 3.7 Sonnet. Change the thinking mode from Normal to Extended, enter and submit your request.
