Anthropic’s latest Claude model delivers responses that are both immediate and thoughtfully considered.

      Anthropic

      OpenAI's o3 and DeepSeek's R1 models are facing new competition. On Monday, Anthropic unveiled its latest "hybrid reasoning" model, Claude 3.7 Sonnet.

      Current reasoning models like o3, R1, and Google’s Gemini 2.0 Flash Thinking are engineered to decompose intricate problems into manageable tasks, subsequently deducing and verifying their findings prior to offering a response. This method yields more precise answers but requires greater computational power and longer inference times. In contrast, Claude 3.7 Sonnet can deliver either "near-instant responses or detailed, step-by-step reasoning visible to the user," as stated in the company’s announcement.

      Claude 3.7’s dual functionality aims to enhance the user experience and remove the cumbersome model selection menus commonly found in other chatbot platforms. OpenAI has also revealed a similar strategy for its forthcoming GPT-4.5 and GPT-5 models.

      “Just as humans utilize a single brain for both quick responses and in-depth reflection, we believe that reasoning should be an integrated capability of cutting-edge models rather than a standalone model,” the company noted. “This cohesive approach also allows for a smoother experience for users.”

      Claude 3.7 Sonnet is being rolled out to all Claude users; however, the model’s extended reasoning feature will require a paid subscription. Anthropic emphasizes that even with its standard reasoning process, Claude 3.7 surpasses its predecessor, Claude 3.5.

      The new Sonnet’s improved reasoning system has been demonstrated to enhance the model’s response quality across various tasks, including math, physics, instruction-following, and coding. “Once again, Claude stands out as best-in-class for real-world coding tasks, with substantial advancements in handling complex codebases and utilizing advanced tools,” the company proudly asserted.

      In the announcement, Anthropic also introduced its agentic AI, named Claude Code. “Claude Code acts as an active collaborator capable of searching and reading code, modifying files, writing and conducting tests, committing and pushing code to GitHub, and utilizing command line tools,” the company stated.

      Claude Code is being released as a limited research preview, with plans for further refinement based on feedback from developers and early users. This agentic AI builds upon the success of Anthropic’s earlier pseudo-agent, Claude Computer Use, which allowed the AI to manipulate its local computing environment by imitating human keyboard and mouse actions.

      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

      

       Andrew Tarantola is a journalist with over a decade of experience reporting on emerging technologies, from robotics to machine…

      

      

      

      

       M4 MacBook Air performance nearly matches Pro models

      

       The M4 MacBook Air is expected to be Apple’s next lightweight and speedy laptop, and a recent benchmark indicates that it may show slight performance enhancements over the latest MacBook Pro.

      Apple is anticipated to announce the new laptop in both 13-inch and 15-inch versions around March. A device believed to be the M4 MacBook Air appeared on the Geekbench 6 benchmark on February 19, revealing results that experts are comparing to other Apple products, including the M4 MacBook Pro and the 13-inch M4 iPad Pro.

      

       Read more

      

      

      

       Users to verify information on Instagram and Facebook. Let's hope it’s effective

      

       A month ago, Meta announced the termination of its third-party fact-checking initiative and revealed it would depend on users to flag false information. Today, the company introduced Community Notes, a similar concept to X, which will be rolled out across Facebook, Instagram, and Threads.

      To qualify as a Community Notes writer, one must be a U.S. citizen aged 18 or older, possess a verified phone number, and have an account older than 6 months. However, not all community notes will be published, as they require a sort of consensus among contributing users.

      

       Read more

      

      

      

       Utilize Microsoft’s Copilot with this Surface Pro 11 sale

      

       If you seek a device to fully leverage Copilot, Microsoft's powerful AI assistant, the Microsoft Surface Pro 11 is a fitting choice. Purchasing it today allows you to benefit from Best Buy's offer of a $350 discount on a package that includes the 2-in-1 laptop with the Surface Pro Keyboard. Instead of the regular price of $1,350, you’ll only pay $1,000, a very reasonable price for such a high-caliber machine. It’s wise not to delay your decision, as these savings may vanish tomorrow.

      Reasons to purchase the Microsoft Surface Pro 11

      The Microsoft Surface Pro 11, featured in our roundup of the best 2-in-1 laptops, is a Copilot+ compatible device. It promises quicker response times and better security for Microsoft's Copilot, enhancing