Anthropic is making a strategic pivot toward life sciences, launching Claude Science as a dedicated AI product for drug discovery and molecular research, while simultaneously renegotiating its pricing agreements with Amazon to boost profit margins ahead of an anticipated public offering later this year. The move signals the company's shift from pure growth expansion into aggressive monetization across both new market segments and existing partnerships.

What Is Claude Science and How Does It Work?

On Tuesday, Anthropic announced Claude Science, a new product designed specifically for scientists and pharmaceutical companies working on drug discovery and molecular biology research. Unlike general-purpose AI tools, Claude Science is built to autonomously execute meaningful scientific work when given high-level instructions, similar to how Claude Code assists software engineers.

The tool includes specialized capabilities for computational biology and drug development, such as 3D protein structure rendering and the ability to interface with genetics, chemistry, and protein biology software commonly used in pharmaceutical research. During a Tuesday demonstration, Alexander Tarashansky, who led Claude Science's development, showed how the system could autonomously identify new drug candidates for phenylketonuria, a rare genetic disease.

Claude Science is now available to all paid Claude subscribers globally, including individual paying users and enterprise subscription customers. The product represents Anthropic's first major offering specifically designed for the life sciences sector, elevating it to the same strategic importance as Claude Code and Claude Cowork, the company's other flagship agent-focused products.

How Does Claude Science Differ From Previous Life Sciences Tools?

Anthropic had previously released plugins and tools for life sciences research in October, but Claude Science marks a significant upgrade in scope and functionality. Rather than being a collection of add-ons, Claude Science is a full-featured, standalone product with deep integration into scientific workflows.

One key distinction is Claude Science's focus on reproducibility and workflow management. The tool not only writes code but also helps scientists run their code on powerful computing clusters, which many researchers need for their work but find difficult to manage independently. This practical focus on infrastructure and reproducibility sets it apart from general-purpose AI assistants.

"We believe that by working in the scientific fields, particularly life sciences and healthcare, we can have a scalable, positive impact on humanity," said Eric Kauderer-Abrams.

Eric Kauderer-Abrams, Head of Life Sciences at Anthropic

Anthropic is not simply providing the tool to external users; the company is also conducting its own drug discovery research using Claude Science, focusing on treatments for rare and neglected diseases. This dual approach allows Anthropic to both advance scientific knowledge and gain real-world insights into how the product performs in practice.

Which Companies Are Already Using Claude for Drug Discovery?

Several major pharmaceutical companies have already integrated Claude into their research and development workflows. Anthropic has signed cooperation agreements with multiple pharmaceutical firms, demonstrating significant market demand for AI-assisted drug discovery. Specific examples include:

• Novo Nordisk: The company has applied Claude in drug discovery, clinical documentation, and regulatory filings, leveraging the AI's ability to process complex scientific and administrative tasks.
• AstraZeneca: The pharmaceutical giant uses Claude to scale its research and development efforts, indicating adoption at the enterprise level.
• Coefficient Bio: Anthropic acquired this biotech startup in April 2026, signaling the company's commitment to building internal expertise in AI-driven drug discovery.

These partnerships underscore the pharmaceutical industry's growing reliance on AI for accelerating the early stages of drug discovery, including molecular design and candidate identification.

How Is Anthropic Changing Its Pricing Model With Amazon?

While expanding into new markets, Anthropic is simultaneously renegotiating its core partnerships to improve profitability. According to reporting from The Information, Anthropic has recently renegotiated agreements with Amazon, switching Claude's pricing model from compute-hours to token-based billing. This change is expected to increase Amazon's costs for using Claude in shopping, programming, and office AI products.

The shift reflects Anthropic's growing bargaining power and its focus on maximizing profit margins as it prepares for an initial public offering. Anthropic's inference infrastructure gross margin has surpassed 70 percent, indicating strong unit economics even as the company scales. The token-based pricing adjustment sends a clear signal to the market that Anthropic is prioritizing profitability over volume growth.

What Are the Key Features of Claude Sonnet 5?

Alongside the Claude Science launch, Anthropic released Claude Sonnet 5, the latest version of its mid-tier AI model. The new model is designed to bridge the gap between high-end flagship models and lightweight options, offering enterprise customers a balance between performance and cost.

Claude Sonnet 5 introduces significant improvements in several areas:

• Agentic Capabilities: The model can plan multi-step tasks, use tools such as web browsers and computer terminals, and complete workflows with minimal human involvement, capabilities previously limited to larger and more expensive models.
• Coding and Reasoning: Internal benchmarks show noticeable improvements over Sonnet 4.6 in reasoning, coding, tool use, and general knowledge work, with better performance on coding evaluations including SWE-bench Pro and Terminal-Bench.
• Computer Interaction: The model demonstrates improved ability to interact with computer interfaces and execute commands, expanding its utility for automation and research tasks.

Anthropic claims that Sonnet 5's agent capabilities are approaching those of its flagship Opus 4.8 model but at a significantly lower cost. The model is now the default AI for users on both Free and Pro plans, while also being available to Max, Team, and Enterprise subscribers.

How to Access Claude Sonnet 5 and Claude Science

• API Access: Developers can integrate Claude Sonnet 5 through the Claude API with introductory pricing of $2 per million input tokens and $10 per million output tokens through August 31, 2026, after which prices increase to $3 and $15 respectively.
• Platform Availability: Both Claude Sonnet 5 and Claude Science are available on Amazon Bedrock, Google Cloud, and Microsoft Foundry, with the latter two currently in preview status.
• User Access: Claude Science is available to all paid Claude subscribers globally, while Sonnet 5 is the default model for Free and Pro users and accessible to Max, Team, and Enterprise subscribers.

How Does Anthropic Plan to Address Biosecurity Risks?

As Anthropic expands its life sciences offerings, the company is also grappling with biosecurity concerns. CEO Dario Amodei has warned that bad actors could potentially use advanced AI models to create new pathogens or bioweapons, and he has emphasized that societal awareness of AI's risks in biology remains low compared to cybersecurity awareness.

To mitigate these risks, Anthropic plans to establish a trusted access program for biological applications, inspired by practices in the cybersecurity sector. The company proposes that only individuals or organizations with verified credentials should be allowed to access powerful bio-AI models with potentially dangerous applications. Notably, Claude Science currently operates on existing models and does not possess advanced biological capabilities; it is designed to assist with drug discovery and molecular research rather than pathogen creation.

"Inside pharmaceutical companies, people routinely handle dangerous biological materials and have their own operating procedures. We can learn from these practices, perhaps we shouldn't open access to anyone before verifying their identity," stated Dario Amodei.

Dario Amodei, CEO of Anthropic

Why Does This Matter for Anthropic's IPO Plans?

Anthropic is at a critical pre-listing stage, with reports suggesting the company could go public as soon as 2026, potentially at a valuation exceeding $1 trillion. The company closed a funding round last month with a pre-closing valuation of $900 billion. The simultaneous launch of Claude Science and renegotiation of Amazon pricing reflects Anthropic's dual focus on expanding revenue streams while improving profitability, both essential metrics for a successful public offering.

The pharmaceutical sector represents a lucrative market opportunity. Pharmaceutical companies have significantly deeper pockets than academic researchers, and major new contracts with these firms could help ensure Anthropic remains profitable as the broader AI market matures. This strategic positioning in life sciences, combined with improved margins from existing partnerships, strengthens Anthropic's financial narrative ahead of its anticipated debut on the capital markets.

Anthropic faces competition in the scientific AI space from OpenAI, which launched GPT-Rosalind in April 2026, a model designed for biological research, drug discovery, and translational medicine. However, John Jumper, a Nobel Prize-winning researcher, recently announced his departure from Google DeepMind to join Anthropic, signaling the company's serious commitment to becoming a leader in AI for science.