Anthropic launches Claude Science – an AI workbench for scientists, not a new model
Anthropic has introduced Claude Science, an AI workbench that provides scientists with a unified environment for computational research, eliminating the need to switch between databases and tools. The platform is not a new AI model but a workflow solution.
Anthropic announced Claude Science on Tuesday, an AI workbench designed to give scientists a single environment for computational research, sparing them the hassle of bouncing between databases, pipelines, and tools. The company emphasizes that Claude Science is not a new AI model and not a more capable model for biology – it runs the same Claude models already available to everyone today (including Claude Opus 4.8), with no special access or gating.
The workbench builds on Anthropic’s October 2025 launch of Claude for Life Sciences, which augmented the Claude chatbot for life sciences tasks. Claude Science is a dedicated place to do that work. The launch fits into Anthropic’s broader strategy to be more than a model provider and to own the operating layer for specific industries, much like Claude Code became the operating layer for software development. The company is increasingly betting on vertical, workflow-level products over raw model capability.
How it works: A main AI assistant acts as a project manager for scientists. It connects to over 60 scientific databases and comes with prebuilt toolkits for fields like genomics, protein structure, and chemistry. This assistant can create sub-assistants to split up work or hand it off to a custom “expert” assistant built by the user. A separate fact-checker AI double-checks citations and calculations before publication, though it is still the same underlying model checking itself.
Claude Science ensures reproducibility by generating figures (e.g., 3D protein structures) alongside the code that created them. Each figure includes the exact code and environment, a plain-language description, and full message history. Scientists can edit figures using plain language, prompting the agent to edit its underlying code. The platform can also run on the lab’s own infrastructure rather than sending data to Anthropic’s servers.
Early users report success. Sean Whalen, a principal scientist at Gladstone Institutes, used Claude Science to build a genome browser from scratch in days. Allen Institute neuroscientist Jérôme Lecoq built a multi-agent computational review pipeline.
The launch follows OpenAI’s GPT-Rosalind in April – a specialized model for biological reasoning. Differences include access: Rosalind was limited to qualified U.S. enterprise customers. Google DeepMind takes a different approach with proprietary models like AlphaFold and AlphaGenome, integrated into Gemini for Science with 30+ life science databases.
Thus, three strategies compete: Anthropic goes wide with broad subscription access; OpenAI goes narrow and enterprise-gated; Google relies on owned, proprietary models. This could signal future competition in other verticals like law, finance, and engineering.
Claude Science is available in beta to Pro, Max, Team, and Enterprise subscribers. Anthropic will support up to 50 projects with up to $30,000 in credits. Applications are open through July 15, 2026.
