Anthropic opens 10,000 free and discounted Claude seats for scientists
Anthropic's new Claude Team plan gives verified researchers free or $15/month premium seats, expands science credits to $50k, and enrolls first Mythos users.
On 27 August 2026, Anthropic announced three things together: a new Claude Team plan for scientists offering free and heavily discounted subscription seats to verified researchers, an expansion of its AI for Science credit programme beyond biology to cover any scientific discipline, and confirmation that a US-government-linked access programme for its most capable Mythos-class models has enrolled its first participants in life sciences research.
Each piece stands on its own. Together, they signal that Anthropic is moving from positioning Claude as a general-purpose assistant to building it as infrastructure that research labs actually depend on.
What the Claude Team plan for scientists offers
The plan opens 10,000 seats to scientists at accredited universities and nonprofit research institutes. Standard seats are free for 12 months. Premium seats, which come with 5× the standard usage limits, are priced at $15 per month for the same period. The pricing is locked in from 27 August 2026, or from the date a group signs up, whichever is later.
Eligibility is straightforward but specific. You need to be a principal investigator (or equivalent) at an academic or nonprofit institution. You complete a verification form, and once approved, you can add the researchers in your lab to the plan. Fields covered include the natural sciences, mathematics, computer science, engineering, and related areas.
If you were already on Anthropic’s earlier Team plan for research labs, your account moves to the new plan automatically. There is one inactivity clause worth knowing: if a group shows no qualifying usage for 90 consecutive days, Anthropic may return it to standard Team pricing so the promotional seats can go to active researchers. They commit to emailing the admin before that happens.
Anthropic says it intends to expand well beyond the initial 10,000 seats over the coming months.
What this means for researchers
Free Claude access for a year is useful on its own terms. The premium tier at $15 per month is a reasonable ask for a lab that needs heavier usage and is already benefiting from the programme. For principal investigators running groups on tight budgets, the ability to bring your whole lab onto a shared plan without negotiating enterprise contracts is genuinely practical.
The 12-month pricing lock also matters. Researchers plan experiments in cycles that often run six months to a year. Knowing the cost is fixed for that period makes it easier to build Claude into a methodology without worrying about mid-project price changes.
One thing worth noting: Anthropic does not train on users’ conversations, uploaded files, or research data under this plan. For researchers handling sensitive or unpublished data, that is an important assurance to have in writing.
The AI for Science credit programme is getting bigger
Separately, Anthropic is expanding its AI for Science programme, which offers API credits for compute-heavy research projects. The programme launched in May 2025 with a focus on biology and life sciences. It has since run several rounds, including a Claude Science-specific round in mid-2026 that funded up to 50 projects at up to $30,000 in credits each, and a rare-disease-focused track in late July and early August 2026 that offered up to $50,000 per team.
The 27 August announcement broadens the scope beyond biology to any scientific field, with a ceiling of $50,000 in credits per project. Anthropic points to work on protein design and progress on the Riemann zeta function as examples of the kind of ambitious, compute-intensive research the programme is designed to support.
If your research sits outside biology and you have previously found the programme out of reach, that exclusion has now been removed. The programme is application-based and reviewed on merit. Groups on the Team plan can apply for extra credits on top of their subscription to support sustained work over several months.
Mythos-class models and the government-linked life sciences programme
The third element of the announcement is more restricted in scope but significant in what it represents. Anthropic is working with the US government to open access to its Mythos-class models for life sciences research and development. The company confirmed on 27 August 2026 that it has enrolled its first participants and expects to expand access soon.
Mythos-class models sit above the Opus class in capability. The first, Claude Mythos Preview, became available in April 2026 through Project Glasswing. Access has remained tightly controlled. Anthropic’s relationship with the US government on this front began with a Department of Energy partnership in December 2025, followed by a Commerce Department agreement that gave select organisations access to Mythos models under specific restrictions.
For the life sciences specifically, Anthropic’s plan is to provide access to Fable 5 with biology and chemistry safeguards removed, while keeping cyber safeguards in place. This matters because Fable 5, under standard conditions, blocks professional biology and drug development queries. Anthropic noted in its August 2026 safeguards update that Fable 5 can outperform experts on some complex biological tasks and could provide meaningful uplift to a bad actor. The trusted access programme is designed to give vetted life sciences organisations the benefit of those capabilities within a controlled framework.
If you work in biomedical research or translational science and are affiliated with an institution that might qualify, it is worth watching for enrolment updates. Anthropic says it will share more details and increase access in the near term.
A note on the Model Hardware Standard
The announcement also includes one item that sits slightly apart from the others. Anthropic is opening a research preview of the Model Hardware Standard (MHS), a shared specification for AI agents to safely operate physical laboratory devices, to a first group of scientific research labs and advanced manufacturers.
MHS grew out of a collaboration between Anthropic’s Beneficial Deployments team and a postdoctoral scientist at HHMI Janelia Research Campus, who was running complex brain-imaging experiments on a rig combining lasers, motorised focusers, and cameras from different vendors with no common interface. The standard is designed to let AI models interact with physical instruments across vendors through a common layer. It is early-stage and limited to a small preview group, but it points toward a future where AI is not just analysing data from experiments but actively operating the equipment that runs them.
The broader picture
Anthropic launched Claude Science on 30 June 2026 as a workbench integrating more than 60 curated tools and connectors pre-configured for fields including genomics, single-cell analysis, proteomics, structural biology, and cheminformatics. The 27 August announcements are built on top of that foundation. The Team plan gives researchers affordable access to the platform. The credit programme gives compute-heavy projects the runway to go further. The Mythos programme gives a carefully vetted tier of life sciences researchers access to the most capable models available.
For a researcher who has been watching AI tools from a distance, wondering whether the cost and complexity were worth it, this is the moment where the barrier gets meaningfully lower.