AI company Anthropic has rolled out a web search function for its assistant, Claude, enabling it to pull real-time information from the internet. The feature, currently in preview for paid users in the United States, is designed to improve response accuracy on topics requiring recent data.
Anthropic states that the search function allows Claude to cite direct sources, giving users the ability to verify information. Instead of manually searching online, users receive summarised insights in a conversational format.
The feature is expected to be applied across multiple industries. As per the company, sales teams can monitor market trends, financial analysts can track earnings reports, and researchers can access primary sources for studies. Consumers can also compare product details and reviews before making purchases.
At present, the web search integration is limited to paid users in the United States, with future plans to extend access to free users and other regions. Users can activate the feature by adjusting their profile settings before using Claude 3.7 Sonnet.
This development aligns with a wider trend of incorporating live data access into AI assistants, as companies refine tools for real-time information retrieval.
Last month, Anthropic launched its latest language model, Claude 3.7 Sonnet, taking on the likes of ChatGPT and DeepSeek. Anthropic also took its first steps towards AI agents with a new tool for developers called Claude Code.
Notably, Claude moved away from the traditional narrative of having a pre-trained model and an independent reasoning model. Instead, the Dario Amodei-led company has integrated the reasoning model alongside the frontier model, meaning that users can simply ask the chatbot when they want it to think longer before answering, essentially triggering the reasoning mode where it reflects on itself before answering a query.
Claude said that triggering the reasoning mode (or as they call it, “extended thinking mode”) helps to improve the model’s performance in maths, physics, following instructions and coding, among many other tasks.