Up to 10,000 researchers can now run Claude without paying. Anthropic’s new scientist plan, announced August 27, gives verified academics free Team seats, with premium seats that multiply usage limits fivefold priced at $15 a month for the first year.
The program targets principal investigators and equivalents at accredited universities and nonprofit institutes, who must pass a verification form before adding lab members. Groups range from 1 to 25 seats. Standard plans stay off-limits to for-profit companies, contract research organizations and industry R&D teams, which are pointed to normal Team and Enterprise pricing. Conversations, files and research data are not used for training, Anthropic says, and accounts idle for 90 days fall back to standard rates.
The wider AI for Science credit program now funds compute-heavy projects outside the biological sciences, with up to $50K in credits per project. Anthropic points to two results. One is on the Riemann zeta function, where an unreleased research version of Claude raised a longstanding bound from 41.6 percent to 67.2 percent. The other is protein design, where its models built binders against 15 targets and succeeded on 14.
Life sciences work stays limited to Opus-class models, with Fable models still blocking biology and drug development queries over dual-use risks. A US government partnership is enrolling researchers for access to Mythos-class models.
