Anthropic is expanding Claude’s role in science with a new Claude Team plan for scientists, offering 10,000 seats for one year and free standard seats, plus premium seats at a locked price of $15 per month for higher usage.

Standard Claude Team seats are free, while premium seats come with higher usage limits at $15/month for one year, and eligibility requires verification of accredited academic or nonprofit research institutions, with lab leaders able to invite team members.

Researchers can apply for up to $50,000 in credits under the AI for Science program, which has supported breakthroughs in fields from protein design to mathematics like the Riemann zeta function.

A government-partnered Mythos program targets life sciences domains such as virology, toxicology, molecular design, and drug development, addressing dual-use risks where commercial safeguards may fall short.

Strategically, the move positions Claude as foundational for AI-assisted science, emphasizing environments where rigorous researchers test models to convert capabilities into reliable knowledge.

The rollout raises equity questions since eligibility hinges on PI affiliation with accredited institutions, potentially limiting independent researchers, though the grant program extends beyond biology to math, physics, chemistry, and related fields.

Safety classifiers explain access gating for frontier biology capabilities, with policy reductions and government pathways shaping who can access the most capable biology models.

The expansion signals a broader trend of AI firms courting academia to drive adoption and safety leadership, potentially accelerating breakthroughs across biology, physics, chemistry, and more.

Other announcements include a research preview of the Model Hardware Standard for safe operation of physical devices, 5 million in wellbeing AI grants, and a clarification article on Claude’s text watermarking.

Anthropic asserts it does not train on users’ conversations, uploaded files, or research data.

In addition, no training uses the content of conversations or uploaded data, reinforcing user data protections.

Recent science results bolster the case for frontier AI in science, with Claude advancing the Riemann zeta bound and autonomous protein design, while Mythos Preview accelerates biology-focused research.