Anthropic offers $3,850 weekly fellowships

Anthropic has opened applications for a new round of its Fellows Program, a paid research track that places engineers and researchers on four-month AI safety projects with company mentors. The company says fellows will receive $3,850 a week in the United States, plus compute funding of about $15,000 a month and access to shared workspaces in Berkeley or London. The postings appeared on Anthropic’s website this week, alongside a longer blog post on the company’s alignment site describing the next 2026 cohorts. The program is notable in part for how Anthropic frames eligibility. Its main application page says the fellowship is meant to support “promising technical talent — regardless of previous experience,” and the company explicitly encourages people to apply even if they do not meet every listed qualification. Anthropic says it reviews applications on a rolling basis and expects the next cohort to start in late September. (alignment.anthropic.com) ### How much money is Anthropic offering? Anthropic’s careers pages list the expected base stipend at $3,850 per week in the United States, 2,310 pounds in Britain, and 4,300 Canadian dollars in Canada. The company says the fellowship is structured as 40 hours a week for four months, with a possible extension. It also says fellows receive benefits that vary by country. (anthropic.com) The same postings say fellows will get funding for compute and other research expenses of about $15,000 a month. Anthropic says participants will also have direct mentorship from its researchers and access to a broader AI safety and security research community. ### Who can apply if they do not have a Ph.D.? Anthropic does not present the fellowship as a Ph.D.-only program. (anthropic.com) The company says it is providing funding and mentorship to technical talent “regardless of previous experience,” language that broadens the pool beyond conventional academic hiring channels. Its application page also says not all strong candidates will meet every qualification and urges interested people not to exclude themselves prematurely. The 2024 launch post for the fellowship used similar language. That earlier announcement said Anthropic wanted to support technical talent “regardless of their level of previous experience” as they made their first move into AI safety research. ### What kind of work would fellows do? Anthropic’s 2026 program post says fellows spend four months on empirical research questions aligned with the company’s priorities, with the aim of producing a public output such as a paper. (anthropic.com) The listed areas include scalable oversight, adversarial robustness and AI control, model organisms, mechanistic interpretability, AI security and model welfare. Anthropic says mentors propose project ideas and fellows help shape them. (alignment.anthropic.com) The company also points to examples from earlier cohorts. Anthropic says more than 80% of fellows in its first cohort produced papers, and more than 40% later joined the company full-time. It says past fellows worked on projects including agentic misalignment, subliminal learning, rapid responses to ASL3 jailbreaks and open-source circuits. (alignment.anthropic.com) ### How does this compare with Anthropic’s earlier fellowship? Anthropic’s 2024 pilot looked smaller and cheaper than the current version. The launch post for that program described a cohort of 10 to 15 fellows working for six months, with a stipend of $2,100 a week and about $10,000 per fellow per month for research costs. The updated 2026 materials describe a four-month program with a higher weekly stipend and higher compute support. (alignment.anthropic.com) Jan Leike, who was quoted in both the older and newer fellowship materials, called the program “an exceptional opportunity to join AI safety research” in Anthropic’s alignment posts. Anthropic’s latest blog says it is opening applications for cohorts beginning in May and July 2026, while its main careers page says applications are rolling for cohorts starting in July 2026 and beyond, with the next cohort expected to begin in late September. (alignment.anthropic.com) ### Where do applicants go next? Anthropic’s application links are live on its careers site now, including a main Fellows Program posting and workstream-specific listings such as AI Security. The company says interviews include an initial application and reference check, technical assessments and interviews, and a research discussion. For applicants who miss the earlier 2026 windows, Anthropic says it runs multiple cohorts each year and reviews submissions on a rolling basis. (alignment.anthropic.com) (anthropic.com)