Program profile · official sources
Agentic AI Program
Stanford Online's Agentic AI program, from the Stanford School of Engineering, adapts a graduate-level Stanford curriculum taught by Adjunct Professor Chowdhery and Assistant Professor Mirhoseini for working professionals. It is 100% online and on demand, with 90 days of access and about 30 hours of video and coding assignments, covering test-time scaling, verification, self-improvement, tool use, memory, planning and agent evaluation. A short application comes first; Python, LLM APIs and basic machine learning are expected. Completion earns a Stanford Certificate of Completion as a digital badge. As shown in the United States on 4 October 2026, tuition is USD 1,650; OpenRouter tokens for assignments, estimated at about USD 100, are extra. The page states no launch date.
Facts checked 4 Oct 2026
- Price
- USD 1650 (As shown in the United States on 4 October 2026, the official program page displays Tuition $1,650 for an individual enrollment with 90 days of access. Assignments also need 100M tokens bought separately from OpenRouter.ai; the provider estimates about $100 of additional compute expenses. Taxes are not stated.)
- Time
- 90 days access, 30 hours to earn the certificate (provider-listed)
- Level
- Graduate-level curriculum adapted for working professionals (provider wording)
- Credential
- Certificate of Completion
Issued as a digital badge verified on the blockchain, which can be shared with your network and verified by employers.
- Prerequisites
- A short application (about 15 minutes) before enrolling. Python programming experience, basic familiarity with LLMs and LLM APIs, and a basic understanding of machine learning concepts such as those in CS229 / XCS229.
- Format
- 100% online, on-demand course with assignments and two optional live sessions; Stanford School of Engineering professional certificate program
Commercial status: no commercial relationship is documented for this page. The verification link goes directly to the official provider source.
Sources ↗
Chart scrolls horizontally. Open the full-size chart.
See the full registry price landscape. Prices retain their stated payment scope; they are not comparable completion costs.
What’s covered
- Test-Time Scaling
- Self-Improvement: Rewards & Verification
- Self-Improvement Techniques
- Reasoning and Self-Improving Agents
- Agentic Frameworks for SW Engineering & Augmenting LLMs with Memory
- Long Horizon Tasks and Agentic Evaluations
- Next Frontiers in Agentic AI
Prerequisites: A short application (about 15 minutes) before enrolling. Python programming experience, basic familiarity with LLMs and LLM APIs, and a basic understanding of machine learning concepts such as those in CS229 / XCS229.
What you actually do
- Test-Time Scaling : Course outline topic · Source
- Self-Improvement: Rewards & Verification : Course outline topic · Source
- Self-Improvement Techniques : Course outline topic · Source
- Reasoning and Self-Improving Agents : Course outline topic · Source
- Agentic Frameworks for SW Engineering & Augmenting LLMs with Memory : Course outline topic · Source
- Long Horizon Tasks and Agentic Evaluations : Course outline topic · Source
- Next Frontiers in Agentic AI : Course outline topic · Source
What the credential is
- Credential type
- Certificate of Completion · online.stanford.edu / agentic-ai
- Credential
- Stanford Certificate of Completion in Agentic AI · online.stanford.edu / agentic-ai
- Issuer
- Stanford School of Engineering (Stanford Online) · online.stanford.edu / agentic-ai
- Sharing
- Issued as a digital badge verified on the blockchain, which can be shared with your network and verified by employers. · online.stanford.edu / agentic-ai
- Exam requirement
- Successfully complete the course requirements; the page lists course materials, video lectures, assignments and exams. · online.stanford.edu / agentic-ai
Total cost to finish
Estimate for the stated payment scope: USD 1650 × 1 = USD 1650.
- Individual enrollment (90 days of access): USD 1650 × 1 = USD 1650. One individual enrollment as displayed on the official page. · Source
Scope: Displayed tuition plus provider-estimated compute expenses of about $100 for OpenRouter tokens; taxes are not stated. · Source
Who should skip it
- Not included in tuition: the 100M OpenRouter.ai tokens needed for in-course assignments, bought separately; the provider estimates about $100 of additional compute expenses. · Official source · 2026-10-04
- Not for learners without Python, LLM API and basic machine learning experience: the provider asks you to enter with a solid foundation in all three, and enrollment starts with a short application. · Official source · 2026-10-04
Change log
- : Initial record from the official program page captured from the United States on 4 October 2026. · Source