AI certifications now come from three kinds of issuers: the cloud platforms (Google Cloud, AWS, Microsoft Azure), the AI infrastructure and data companies (NVIDIA, Databricks), and the AI labs (Anthropic, with OpenAI in a pilot). They differ a lot in cost, length, who can sit them and what they prove.
This guide lists the real, proctored certifications we could confirm on official pages, as of October 2026. For each one you get who it fits, the price, the exam format and any stated prerequisites. Every detail comes from the issuer's own page, linked inline. Where a page does not state a detail, we leave it out.
If you want free learning first, start with our list of the best free AI courses in 2026. A certification makes the most sense after you have some skills to prove.
Certificate or certification?
The two words get mixed up, and the issuers themselves draw the line.
- A certification is a proctored exam with identity checks. You pass a scored test, and you get a credential that usually expires and needs renewal.
- A certificate of completion or a course badge says you finished a course and passed its quiz.
OpenAI's help center is direct about this. Its OpenAI Academy courses are free, and a badge means you completed a course and passed its assessment with a score of 80% or higher. The page says these badges and certificates of completion "are not certifications and do not guarantee eligibility for a future certification." Anthropic's certification eligibility page makes the same split: a Claude Academy certificate of completion records that you finished a course and passed its quiz, while a Claude certification is a proctored exam delivered by Pearson VUE.
Everything below is a certification in the first sense.
Entry level: no coding required
These fit managers, analysts, marketers, students and anyone who wants to show they understand AI concepts without building systems.
Google Cloud Generative AI Leader
Generative AI Leader is Google Cloud's business-level certification. Google says it "is for anyone in any job role, with or without hands-on technical experience."
- Cost: $99 (plus tax where applicable)
- Format: 90 minutes, 50-60 multiple choice questions, online-proctored or onsite-proctored
- Languages: English, Japanese, Spanish, Portuguese
- Prerequisites: none
- Validity: 3 years
- What it covers: fundamentals of gen AI, Google Cloud's gen AI offerings, techniques to improve model output, and business strategies for a gen AI solution
Who it fits: people who lead or influence AI projects at a company that uses Google Cloud, or anyone who wants the lowest-cost entry exam on this list.
AWS Certified AI Practitioner
AWS Certified AI Practitioner is AWS's foundational AI exam. AWS calls it "a good starting point for individuals exploring AI/ML on AWS." The page says the ideal candidate is familiar with AI/ML on AWS "but does not necessarily build AI/ML solutions on AWS."
- Cost: $100
- Format: 90 minutes, 65 questions, at a Pearson VUE testing center or online proctored
- Languages: 12, including English, Japanese, Korean, Simplified Chinese and Spanish. AWS says the Italian and German versions will be retired after October 15, 2026.
- Validity: 3 years. Earning the Machine Learning Engineer Associate certification recertifies it automatically.
- Prerequisites: none stated. AWS suggests people new to IT and AWS take AWS Cloud Practitioner Essentials or AWS Technical Essentials first.
Who it fits: people in cloud, data, IT or business roles at AWS shops, and anyone who plans to move up to the AWS Machine Learning Engineer Associate later.
AWS Certified AI Business Strategist
AWS Certified AI Business Strategist is new and in beta. AWS says it tests "your ability to evaluate AI investments, build business cases, design governance, and scale adoption." It "does not assess AWS services knowledge."
- Cost: $50 at beta pricing. AWS lists the standard exam price as $100.
- Format: 170 minutes, 85 questions, multiple choice and multiple response, at a Pearson testing center or online proctored
- Languages: English, Japanese
- Prerequisites: none required. AWS recommends six months of experience working with or alongside AI initiatives, and says no coding or AWS implementation experience is needed.
- Extra: people who earn it by February 15, 2027 get an additional Early Adopter digital badge.
Who it fits: product and program managers, consultants, business analysts, sales and marketing staff who make AI decisions but do not write code. The beta price makes it the cheapest proctored exam here while it lasts.
Microsoft Certified: Azure AI Fundamentals (Exam AI-901)
Microsoft replaced the old AI-900 exam, according to Pearson VUE's retirement list. Azure AI Fundamentals now requires Exam AI-901, and Microsoft says the English version was updated on April 15, 2026.
Note that this one is less "no code" than its name suggests. Microsoft says candidates need "knowledge of Python coding syntax and programming techniques" and should be familiar with Azure resources.
- Cost: Microsoft does not list one price. The exam page says the price is "based on the country or region in which the exam is proctored."
- Passing score: 700
- What it covers: identify AI concepts and capabilities (40-45%), and implement AI solutions by using Microsoft Foundry (55-60%)
- Languages: 13, including English, Japanese, Korean, Spanish and French
Who it fits: students and early-career developers who want a first Microsoft credential and can already read basic Python.
Builder level: you write code with models
These fit developers, data engineers and ML engineers who build applications on top of models. For a full path into this kind of work, see how to become an AI engineer in 2026.
NVIDIA-Certified Associate: Generative AI LLMs (NCA-GENL)
NVIDIA describes NCA-GENL as "an entry-level credential" for the concepts behind building and maintaining applications with generative AI and LLMs on NVIDIA solutions.
- Cost: $125
- Format: 1 hour, 50-60 multiple-choice questions. You need a Certiverse account to take it.
- Language: English
- Prerequisites: "A basic understanding of generative AI and large language models"
- Validity: two years. You recertify by retaking the exam.
- Topic weights: Core Machine Learning and AI Knowledge 30%, Software Development 24%, Experimentation 22%, Data Analysis 14%, Trustworthy AI 10%
Who it fits: junior developers and data people who want an LLM credential that is not tied to one cloud provider. The topic list includes RAG, embeddings, prompt engineering, fine-tuning data prep and model evaluation, so it tests general LLM skills, not one cloud provider's services. If RAG is new to you, our plain-English RAG explainer covers the basics.
Databricks Certified Generative AI Engineer Associate
The Databricks Generative AI Engineer Associate exam tests whether you can "design and implement LLM-enabled solutions using Databricks." Databricks says people who pass "can be expected to build and deploy performant RAG applications and LLM chains."
- Cost: $200
- Format: 90 minutes, 45 scored multiple choice questions, online or at a test center. No test aids. The exam may include extra unscored items.
- Languages: English, Japanese, Portuguese (Brazil), Korean
- Prerequisites: none, but Databricks says related training is "highly recommended" and recommends 6+ months of hands-on experience with the tasks in the exam guide
- Validity: 2 years
- Topic weights: Design Applications 14%, Data Preparation 14%, Application Development 30%, Assembling and Deploying Apps 22%, Governance 8%, Evaluation and Monitoring 12%
Who it fits: data engineers and ML engineers whose company already runs on Databricks. Machine learning code on the exam is in Python.
Microsoft Certified: Azure AI Apps and Agents Developer Associate (Exam AI-103)
This is Microsoft's new developer-level AI certification. The older Azure AI Engineer Associate (Exam AI-102) now shows a notice that the certification and its renewal assessment are retired. Pearson VUE's Microsoft updates page lists the Azure AI App and Agent Developer Associate (Exam AI-103) as the related new certification.
Azure AI Apps and Agents Developer Associate "validates your expertise in designing, developing, and deploying advanced Azure AI solutions using Python and Microsoft Foundry."
- Cost: based on the country or region where you take the exam
- Format: 120 minutes, proctored, and Microsoft says it may have interactive components
- What it covers: plan and manage an Azure AI solution, implement generative AI and agentic solutions, computer vision, text analysis, and information extraction
- Prerequisites: none stated. Microsoft says you should have experience developing apps in Python.
Who it fits: developers at Microsoft and Azure shops who build agents and AI apps.
AWS Certified Machine Learning Engineer Associate
The AWS Machine Learning Engineer Associate is in the middle of a version change, so check the page before you book.
- Current state: The last day to take the old MLA-C01 exam in English was September 28, 2026. The updated MLA-C02 is in beta, in English only. MLA-C01 stays available in Korean, Japanese and Simplified Chinese until MLA-C02 reaches general availability, which AWS lists as TBD.
- MLA-C02 beta: $75 (beta pricing), 170 minutes, 85 questions, multiple choice and multiple response
- MLA-C01: $150, 130 minutes, 65 questions
- Validity: 3 years
- Recommended experience: at least 1 year of experience in machine learning engineering or a related field, and 1 year of hands-on experience with AWS services
- What changed: AWS says MLA-C02 adds generative AI implementation (Amazon Bedrock, RAG architectures), agentic AI, foundation models and LLMs, and responsible AI practices
Who it fits: ML engineers, MLOps engineers and data scientists on AWS. The beta is cheaper than the standard exam, but AWS says beta exams contain additional unscored items and results typically arrive within 5 business days.
Advanced level: production experience expected
Google Cloud Professional Machine Learning Engineer
Professional Machine Learning Engineer covers building, deploying and monitoring traditional and generative AI models on Google Cloud. Google says the exam was updated to reflect the transition from Vertex AI to Gemini Enterprise Agent Platform.
- Cost: $200 (plus tax where applicable)
- Format: two hours, 50-60 multiple choice and multiple select questions, online-proctored or at a testing center
- Languages: English, Japanese
- Prerequisites: none
- Recommended experience: 3+ years of industry experience, including 1 or more years designing and managing solutions using Google Cloud
- Coding: Google says the exam "does not directly assess coding skill," but you should have minimum proficiency in Python and SQL to read code snippets
Who it fits: working ML engineers on Google Cloud who want a senior-level credential.
AWS Certified Generative AI Developer Professional
AWS Certified Generative AI Developer Professional targets developers who build production generative AI on AWS services such as Amazon Bedrock.
- Cost: $300
- Format: 180 minutes, 75 questions, multiple choice or multiple response, at a Pearson VUE testing center or online proctored
- Languages: English, Japanese, Korean, Simplified Chinese
- Prerequisites: no certification required. AWS says the target candidate has 2 or more years of experience building production-grade applications on AWS or with open-source technologies, general AI/ML or data engineering experience, and 1 year of hands-on experience implementing generative AI solutions.
Who it fits: senior developers who already ship generative AI features on AWS. It is the most expensive exam on this list.
AI lab certifications
Claude Certification Program (Anthropic)
Anthropic runs the Claude Certification Program, with exams delivered by Pearson. Pearson's Claude Certification Program page says certification "is open to organizations in the Claude Partner Network." Anthropic's certification FAQ says registration needs a partner email address on a recognized company domain, and personal email addresses will not work. The FAQ says certification "is currently available only to organizations in the Claude Partner Network."
The four certifications, with prices in USD before any partner-tier discount, per the FAQ:
- Claude Certified Associate: $99
- Claude Certified Developer: Foundations: $125
- Claude Certified Architect: Foundations: $125
- Claude Certified Architect: Professional: $175
Other details from the FAQ:
- Format: multiple choice and scenario-based multiple response questions, 120 minutes. Closed book.
- Passing score: 720 on a scale of 100 to 1,000, for all four
- Validity: 12 months. On-time renewal is a free, non-proctored assessment.
- Discounts: Select, Preferred and Global Premier partners get 50% off. Through December 31, 2026, Global Premier partners get a 100% discount.
- Languages: English and Japanese
Pearson says the Associate, Developer: Foundations and Architect: Professional exams get updated content starting December 7, 2026, with updated exam guides on November 17, 2026.
Who it fits: consultants and engineers at Claude Partner Network firms. If you work at one of those firms, these are the credentials Anthropic itself issues for Claude work.
OpenAI
OpenAI announced its first OpenAI Certifications courses on December 9, 2025, with AI Foundations launching in ChatGPT through pilot programs with employers and public-sector partners. As of October 2026, the OpenAI Certified app "is available only for ChatGPT Enterprise and Edu workspaces on an invite-only basis." The page tells people interested in enrolling to contact their OpenAI account representative.
What anyone with a ChatGPT account can do is take the free OpenAI Academy courses and earn badges and pathway certificates of completion. OpenAI says those are not certifications.
How to choose
Answer these questions in order. Each one narrows the list.
Do you write code? If not, look only at the entry track: Google Cloud Generative AI Leader, AWS AI Practitioner or AWS AI Business Strategist. Azure AI Fundamentals expects basic Python, so it fits people who can read code.
Which platform does your employer or target employer use? A cloud certification proves the most when the job uses that cloud. Pick Google Cloud, AWS, Azure or Databricks to match. If you do not know yet, look at NVIDIA's NCA-GENL: its topics are general LLM skills such as RAG, embeddings and evaluation, not one cloud provider's services.
How much experience do you have? Match the exam to the experience the issuer recommends. The two advanced exams recommend several years of experience: 3+ years of industry experience, including 1+ year on Google Cloud, for Google's Professional ML Engineer, and 2+ years building production-grade applications plus 1 year of hands-on generative AI work for the AWS professional exam.
What is your budget? Standard prices on this list run from $99 to $300 per attempt, before tax. AWS has two exams at beta prices right now ($50 and $75). Microsoft prices depend on your country, so check at registration. Retakes cost money too: AWS says you pay the full registration fee for each attempt, and Anthropic says a retake costs the full exam fee.
How long will it stay valid? Validity ranges from 12 months (Claude) to 2 years (NVIDIA, Databricks) to 3 years (AWS AI Practitioner, AWS Machine Learning Engineer Associate, Google Cloud Generative AI Leader). A shorter window means more renewal work.
Does a certification beat a portfolio? A certification shows you passed one exam. Pair it with projects you can show. Our look at entry-level AI jobs in 2026 covers what else helps junior candidates get hired.
Before you book
- Read the exam guide linked from each official page. It is the authoritative list of topics.
- Check the exam version. AWS ML Engineer Associate is between versions, Microsoft retired AI-900 and AI-102, and Anthropic updates three exams on December 7, 2026.
- Use a personal account where the issuer suggests it. Microsoft "strongly" recommends registering with a personal Microsoft account, because exam records on a work or school account are lost if you leave.
- Prices on this page are as of October 2026 and in USD unless noted. Confirm the price at checkout.
Sources
- Google Cloud: Generative AI Leader
- Google Cloud: Professional Machine Learning Engineer
- AWS Certified AI Practitioner
- AWS Certified AI Business Strategist
- AWS Certified Machine Learning Engineer Associate
- AWS Certified Generative AI Developer Professional
- AWS Certification FAQs
- Microsoft Certified: Azure AI Fundamentals
- Microsoft Exam AI-901
- Microsoft Certified: Azure AI Apps and Agents Developer Associate
- Microsoft Certified: Azure AI Engineer Associate (retired)
- Pearson VUE: Microsoft exam updates
- NVIDIA: Generative AI LLMs certification (NCA-GENL)
- Databricks Certified Generative AI Engineer Associate
- Anthropic: Four role-based Claude certifications
- Anthropic Partner Academy: Certification FAQ
- Anthropic Partner Academy: Who can take a Claude certification exam
- Pearson VUE: Claude Certification Program
- OpenAI: Launching our first OpenAI Certifications courses
- OpenAI Help Center: OpenAI Certified app
- OpenAI Help Center: OpenAI Academy courses
