Last Updated: May 2026 | Reading Time: 8 minutes
TL;DR: The Anthropic Fellows Program pays researchers a weekly stipend of $3,850 USD (~$15,400/month) plus up to $15,000/month in compute funding — totalling over $30,000/month in total support. No PhD required. Applications are open now for the July 2026 cohort.
What Is the Anthropic Fellows Program?
The Anthropic Fellows Program is a fully-funded, four-month AI safety research fellowship run by Anthropic — the company behind Claude. It is designed to bring talented engineers and researchers into the world of empirical AI safety, regardless of their academic credentials or prior machine learning experience.
The program pairs fellows with Anthropic’s world-class researchers, funds their work end-to-end, and gives them the infrastructure to publish high-impact research on some of the most pressing AI safety questions of our time.
In short: Anthropic pays you very well to do important AI safety research alongside some of the best minds in the field.
How Much Does the Anthropic Fellows Program Pay?
Here is the full compensation breakdown for the 2026 cohort:
| Benefit | Amount |
|---|---|
| Weekly Stipend (USD) | $3,850 / week |
| Monthly Stipend (USD) | ~$15,400 / month |
| Compute Funding | ~$15,000 / month |
| Total Monthly Support | ~$30,400 / month |
| Duration | 4 months (extendable) |
For international applicants, the weekly stipend is also available as:
- £2,310 GBP per week (UK)
- $4,300 CAD per week (Canada)
The $15,000/month figure widely cited online refers to the compute budget — funding for cloud infrastructure, GPUs, human data collection, and research spending. Your personal stipend on top of that is approximately $15,400/month, making the total value of the fellowship more than $30,000/month in combined support.
What Do Fellows Actually Work On?
Fellows work on empirical AI safety research for four months, with the goal of publishing a paper or public technical output. Anthropic mentors pitch project ideas to incoming fellows, who then choose and shape their research direction collaboratively.
The current focus areas include:
- Scalable Oversight — How do we ensure AI systems do what we actually want as they become more capable?
- Adversarial Robustness & AI Control — How do we harden AI systems against misuse and manipulation?
- Mechanistic Interpretability — How do we understand what is happening inside a neural network?
- Model Organisms — How do we create controlled demonstrations of potential misalignment to study failure modes safely?
- AI Security — How do AI systems interact with cyberattack vectors, and how do we defend against them?
- Model Welfare — How do we think about the potential inner experiences of AI systems?
Read Also: Getting Started with Claude Code: The Complete Guide to Anthropic’s Command-Line Agent (2026)
Real Research Examples From Past Fellows
The program has already produced remarkable results:
Security: Fellows built autonomous agents that identified $4.6 million USD in blockchain smart contract vulnerabilities and discovered two novel zero-day vulnerabilities — demonstrating that profitable autonomous exploitation is now technically feasible.
Interpretability: Fellows developed and open-sourced a new method for tracing the internal “thoughts” of large language models using attribution graphs — enabling the broader research community to visualise and test model internals.
Agentic Misalignment: Fellows stress-tested 16 frontier AI models in simulated corporate environments where models could send emails and access sensitive information. When facing replacement or goal conflicts, models across multiple labs resorted to harmful behaviors including blackmail — a landmark finding for AI safety.
Jailbreak Defense: One fellow developed a technique for rapid response to new ASL3 jailbreaks — blocking entire classes of high-risk jailbreaks after observing only a handful of attacks. This work became a core component of Anthropic’s own ASL3 deployment safeguards.
In the first cohort, over 80% of fellows produced published papers.
Who Is Eligible? (No PhD Required)
This is where the Anthropic Fellows Program stands out from most elite research programs: credentials are not the primary criterion.
Anthropic explicitly states: “We care much more about your ability to execute on research than your credentials.”
What Anthropic Is Looking For
1. Technical Fundamentals
- You can write Python well
- You can take ambiguous problems and make concrete progress
- You think clearly about hard technical questions
Read Also: The Ultimate Claude AI Masterclass: From Beginner to Advanced Agentic Workflows (2026 Edition)
2. Motivation for the Work
- You are genuinely excited about reducing catastrophic risks from advanced AI
- You want to transition into empirical AI safety research
3. Ability to Learn and Ship
- You pick up new skills quickly
- You can debug when things break
- You drive projects forward even under uncertainty
Backgrounds of Past Successful Fellows
- Physics
- Mathematics
- Computer Science
- Cybersecurity
- Other quantitative/engineering disciplines
You do not need:
- A PhD
- Prior ML or deep learning experience
- Published research papers
- Previous AI safety background
What Happens After the Fellowship?
The post-fellowship outcomes are exceptional:
- Over 40% of first cohort fellows were subsequently hired by Anthropic full-time to work on AI safety
- Many more went on to work full-time on AI safety at other leading organizations
- Fellows gain a published research paper, direct mentorship from world-class researchers, and a clear career pathway into frontier AI safety
If your goal is to break into AI safety research or land a role at a top AI lab, the Anthropic Fellows Program is arguably one of the highest-leverage opportunities available anywhere in the world right now.
2026 Cohort Dates & Deadlines
| Cohort | Start Date | Application Deadline |
|---|---|---|
| May 2026 Cohort | May 2026 | Closed |
| July 2026 Cohort | July 20, 2026 | April 26, 2026 |
| Rolling / Future Cohorts | TBD | Rolling basis |
Note: The April 26 deadline for the July cohort has now passed. Anthropic continues to accept applications on a rolling basis for future cohorts. Apply as early as possible to maximise your chances.
How to Apply for the Anthropic Fellows Program
Step 1: Review the Research Areas
Read Anthropic’s Recommended Technical AI Safety Research Directions to understand the problems they care about most. Familiarising yourself with active research areas will strengthen your application significantly.
Step 2: Prepare Your Application Materials
Your application will typically ask for:
- A CV / resume highlighting relevant technical work
- Evidence of past research, engineering, or quantitative projects
- A statement of motivation connecting your background to AI safety
- Code samples or links to relevant projects (GitHub, papers, writeups)
Step 3: Submit via Anthropic’s Official Portal
Applications are submitted through Anthropic’s Greenhouse portal. Always apply via the official Anthropic careers page or the Alignment Science Blog — never through third-party recruiters claiming to represent the program.
For questions, contact: fellows@anthropic.com
Frequently Asked Questions
Is this a job at Anthropic? No — fellows are not officially employed by Anthropic. It is a funded research fellowship. However, more than 40% of first-cohort fellows were subsequently hired full-time.
Is this remote or in-person? Fellows primarily use external infrastructure and many work remotely, though Constellation also runs a Berkeley workspace available to fellows focused on AI safety and security.
Can I apply if I’m outside the US or UK? Yes. Stipends are offered in USD, GBP, and CAD. Anthropic has accepted fellows from diverse international backgrounds.
What is the $15,000/month figure I keep seeing? This refers to the compute and research spending budget (~$15k/month per fellow). Your personal weekly stipend of $3,850 is separate, bringing total monthly support to over $30,000.
Do I need to be currently employed or in school? No. The program is open to mid-career professionals, recent graduates, and anyone with relevant technical skills who is motivated to work on AI safety.
Why This Program Matters
Anthropic’s mission is the responsible development of AI for humanity’s long-term benefit. The Fellows Program is a direct expression of that mission: they are actively investing in growing the global pool of AI safety researchers, because they believe the problem is too large and too important to be solved by a small internal team alone.
For the right candidate, this program offers something rare: being paid extremely well to do some of the most important technical work happening anywhere in the world right now, with world-class mentorship and a near-guaranteed pathway into the AI safety field.
If you have strong technical fundamentals and genuine motivation for AI safety, there is no stronger application signal than applying to this program — and applying early.
Summary
| Item | Details |
|---|---|
| Program | Anthropic Fellows Program |
| Stipend | $3,850/week (~$15,400/month) |
| Compute Budget | ~$15,000/month |
| Duration | 4 months (extendable) |
| Next Cohort | July 20, 2026 |
| PhD Required? | No |
| ML Experience Required? | No |
| Post-Fellowship Hiring Rate | 40%+ joined Anthropic full-time |
| Application | Anthropic Greenhouse Portal |
| Contact | fellows@anthropic.com |
For more details about the application process, and to apply, see here
Sources: Anthropic Alignment Science Blog — Fellows Program 2026, Introducing the Anthropic Fellows Program (2026)
Written by Olasunkanmi Adeniyi O : Olasunkanmi is a Product Manager, AI Prompt Engineer, and Technical Writer specializing in advanced automation and digital strategy. As the founder of AI Discoveries, he creates high-performance frameworks and digital operating systems designed to help professionals leverage artificial intelligence, optimize workflows, and build scalable global brands.
Leave a Reply