Anthropic’s new AI model cut drug design time 10x

Advertisement

Anthropic launched Claude Fable 5 and Claude Mythos 5, frontier AI models the company says accelerated parts of its internal drug design work roughly tenfold, on June 9.

The two models share the same underlying system — a new “Mythos-class” tier Anthropic positions above its Opus models — and differ only in their safeguards. Fable 5 is available for general use, while the more capable Mythos 5 is restricted to Project Glasswing partners, but will be made available to select biology researchers with some safeguards lifted soon, according to the company’s news release.

In drug design testing, Anthropic reported that Mythos 5, equipped with protein design and bioinformatics tools but no human assistance, matched or beat skilled human operators on tasks normally handled by a scientist, including selecting binding sites and recovering from failed runs. Nine of 14 protein targets in the study yielded strong candidates Anthropic said it is now investigating.

The company also said Mythos 5 is its first model to consistently generate novel scientific hypotheses. In blinded comparisons, its scientists preferred the model’s molecular biology hypotheses about 80% of the time, and one — a proposed mechanism for an E. coli protein — was independently corroborated by an outside lab. In genomics, Mythos 5 worked largely autonomously for more than a week, assembling single-cell data across 138 species and training a model that outperformed a recently published model 100 times its size.

The research capabilities build on the commercial healthcare and life sciences business Anthropic stood up earlier this year. In January, the company rolled out HIPAA-ready Claude for Healthcare — connecting Claude to the CMS coverage database, ICD-10 codes and the National Provider Identifier Registry — and expanded its Claude for Life Sciences platform to draft clinical trial protocols, operate trials and prepare regulatory submissions. Phoenix-based Banner Health was an early enterprise adopter, building an internal AI platform called BannerWise for more than 55,000 employees across six states. That launch came as OpenAI moved into the same market with ChatGPT Health.

Citing dual-use risk, Anthropic routes most biology and chemistry queries on Fable 5 to its Claude Opus 4.8 model. The company plans a trusted-access program giving select life sciences researchers Fable 5 with those safeguards removed. Both models are priced at $10 per million input tokens and $50 per million output tokens.

The Mythos 5 and Fable 5 release occurred just days after the company submitted a draft registration statement on Form S-1 to the U.S. Securities and Exchange Commission for a proposed IPO. The company’s valuation was $965 billion in May after it raised $65 billion in Series H funding.

At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.

Register to Attend Webinar

Designing the intelligent hospital: Building hospitals around people, data and care

Tuesday, July 21
12:00 PM - 1:00 PM CDT

Presenters: Braheem Santos, Schneider ElectricJohn Donohue, Penn Medicine

Advertisement

Next Up in Artificial Intelligence

Advertisement