Anthropic has launched Claude for Life Sciences, a specialized artificial intelligence tool to accelerate scientific research, and expanded its Claude Code agentic coding tool to web and mobile platforms to augment drug discovery and developer workflows.
The introduction of Claude for Life Sciences marks Anthropic’s first targeted entry into the life-sciences sector. The company has customized its AI model to assist scientists and researchers with tasks throughout the drug discovery process. The tool is designed to perform functions such as reading scientific studies, formulating theories based on existing data, analyzing large numerical datasets, and preparing documentation for regulatory submissions. The model has been specifically redesigned to connect with laboratory tools that are commonly used by researchers in their daily work, creating a more integrated environment.
Anthropic’s stated ambition for the new tool was outlined by Eric Kauderer-Abrams, the company’s Head of Biology and Life Sciences. In an interview with CNBC, he described the launch as a significant investment for the company in this particular field. “We want a meaningful percentage of all of the life science work in the world to run on Claude, in the same way that that happens today with coding,” Kauderer-Abrams stated. This objective positions the company’s goals in the life-sciences domain to be as impactful as the role AI has already established in software development and coding assistance.
To facilitate this level of integration, Anthropic has developed a series of direct connections for Claude for Life Sciences. The tool is engineered to work with established research platforms including Benchling, 10x Genomics, PubMed, and Synapse.org. These integrations allow scientists to pull data directly from these platforms into the Claude interface. This capability eliminates the need for researchers to manually export files or switch between different applications, a process that can be time-consuming and introduce potential for error. The streamlined workflow is intended to allow for a more fluid and efficient data analysis process within a single, unified system.
In addition to the technical integrations, Anthropic has formed partnerships with several consulting firms to support the adoption of the new tool. The company is collaborating with Deloitte Touche Ltd., KPMG International Ltd., and Caylent Inc. The role of these partners is to provide guidance and training to life-sciences organizations, helping them to effectively implement Claude for Life Sciences into their existing research and development pipelines and maximize its utility. This collaborative approach is aimed at ensuring a smoother transition for organizations adopting the AI technology.
The development of a dedicated life-sciences tool was reportedly inspired by existing user behavior. Kauderer-Abrams, who joined Anthropic a few months prior to the launch, noted that the company was encouraged by observing that many researchers were already using the standard, general-purpose Claude model to assist with their scientific work. This organic adoption prompted the company to build a specialized version of the model, complete with the necessary infrastructure and integrations tailored specifically for the research community’s needs, creating a more powerful and fit-for-purpose solution.
Claude for Life Sciences is built upon Claude Sonnet 4.5, Anthropic’s most powerful large language model. The underlying architecture has been refined to deliver enhanced performance on specific tasks relevant to the scientific field, such as interpreting complex lab protocols and processing other science-related documents. In a demonstration of its capabilities, the company illustrated a scenario where a scientist conducting preclinical studies used the tool to compare two different dosing plans for a new drug. The demonstration showed the scientist pulling data directly from the Benchling platform, using Claude to auto-generate comparison tables, analyze the differences between the two plans, and then update the information back into the original system.
Following the data analysis phase of the demonstration, the scientist then instructed Claude for Life Sciences to generate a complete study report suitable for submission to regulators. Anthropic noted that this entire sequence of tasks, which would typically require several days of manual work to compile results and validate the findings, was completed within minutes using the AI tool. This highlights the potential for significant time savings in the documentation and reporting stages of research.
Despite the demonstrated efficiency gains, Kauderer-Abrams clarified the limitations of the technology. He pointed out that while AI can accelerate certain aspects of the research process, it cannot speed up the entire drug discovery pipeline. For instance, he specified that it is not possible for artificial intelligence to shorten the duration of clinical trials, which are governed by strict protocols and timelines. The primary benefit of a tool like Claude lies in accelerating the more repetitive and data-intensive components of scientific workflows. These include tasks such as painstaking data analysis, detailed comparisons, and the extensive paperwork required for regulatory compliance. “We’re here to make sure this transformation happens and that it’s done responsibly,” Kauderer-Abrams added, emphasizing the company’s focus on proper implementation.
Concurrent with the life-sciences announcement, Anthropic released an update for Claude Code, its agentic coding tool. Previously available only through a developer’s terminal, the tool is now accessible via the Claude.ai web application and the company’s mobile app. This expansion is being released as a “research preview,” and initial access is limited to users with Pro and Max subscription accounts. This phased rollout allows the company to gather feedback on the new interface from its more advanced user base before a wider release.
The new web and mobile accessibility offers a different way for developers to interact with the coding agent. The interface is designed to be particularly useful for developers who need to run multiple, well-defined coding tasks in parallel. A user can connect the agent to their GitHub repositories, provide instructions for the desired tasks, and then supervise the progress as new code is generated and appears within the web interface. The user interface displays the agent’s activity in a sidebar, with a list of assigned tasks shown on the left-hand side of the screen.
Anthropic has implemented specific security measures for the new Claude Code interface. Each coding session operates within its own sandboxed environment, isolating it from other processes to enhance security. All interactions with Git repositories are routed through a secure proxy service. This proxy is designed to verify authorization, ensuring that the AI agent can only access repositories for which it has been explicitly granted permission, preventing unauthorized code access or modifications.
A new feature called “steering” has also been incorporated into the web and mobile versions of Claude Code. This capability allows a developer to provide guidance to the agent while it is actively working on a problem, without needing to interrupt or stop the ongoing task. This can help refine the agent’s output mid-process and can potentially prevent the need to restart a complex task from the beginning if a deviation occurs. It provides a more interactive and dynamic way to manage the AI’s coding process.
The company described this new operational model as a shift in how developers can leverage AI assistants. “Instead of managing individual coding tasks one at a time, developers can now oversee a fleet of Claude Code instances with confidence they’ll finish safely and independently,” the company stated. It further elaborated on this paradigm: “It’s less about watching Claude work and more about delegating to an entire team — you assign the work, Claude handles the execution, and you review the results when each task completes.”
Anthropic reported that it has been using Claude Code extensively for its own internal development. According to the company, the agent has assisted in writing approximately 90% of its own codebase. The internal use of the tool has reportedly led to a productivity increase of around 67% for its engineering team over the last few months. This internal adoption has also translated into commercial success, with Claude Code generating an annual revenue run rate for the company that now exceeds $500 million.
