Anthropic launched Claude Science on June 30, 2026. It is a new software tool designed to help scientists and researchers do their work more efficiently by connecting to the programs and databases they already use.

Claude Science is built on Anthropic's existing Claude AI models. The company had been moving in this direction for a while. In October 2025, Anthropic released Claude for Life Sciences, a version designed specifically for drug discovery research. Claude Science takes that same approach and expands it to cover all kinds of scientific work.

What Claude Science Does

The tool plugs into the software and databases that researchers are already running. It doesn't ask them to switch to a completely different system. Inside the workbench, researchers can run code to analyze data, create charts, do calculations, and manage files. The output is also important: Claude Science creates a clear record of every step it takes, so researchers can check the work and repeat it later if needed.

Anthropic held an event on June 29, 2026 called "The Briefing: AI for Science," where company leaders and researchers discussed how Claude could help with scientific work. The event happened the day before the product launch — a typical way to build interest with important institutions.

Why This Matters

Scientific research comes in many different forms. A lab studying proteins uses different tools than a lab studying how chemicals work, which uses different tools than one looking at patient records. Most of these don't talk to each other. Claude Science tries to work with all of them at once, without forcing labs to change what they already have in place.

The ability to show its work is the most important feature for institutions that want to use it. Scientific discoveries have to be repeatable — other researchers need to be able to follow the same steps and get the same results. A tool that can explain what it did, step by step, is different from one that just gives you a final answer with no way to check how it got there.

Anthropic is now competing directly with other companies that serve scientists. There are specialized software firms like Schrödinger and Benchling, and giant cloud companies like Amazon, Google, and Microsoft all offer AI tools for research. All of them have more money and more history than Anthropic. The bet Anthropic is making is that Claude thinks through complicated scientific problems better than the other options, and that this tool makes it easier to use — enough so that research teams will switch.

The life sciences product from October 2025 was a smart first step. Drug discovery is expensive, and there are clear ways to measure whether an AI tool saves money and time. It is the kind of place where researchers already expect to use AI for screening compounds. Claude Science now extends that to all research, though researchers and institutions will take longer to decide whether to switch compared to a drug company with a clear budget and deadline.

Anthropic is moving fast to get established in AI for science before the market settles into a few winners. The company chose to build something that works alongside existing research tools rather than replace them — a choice that reflects how research institutions actually work. They have strict rules about data and security, and switching everything to a new platform is rarely possible or practical.