OCI AI Foundations Certification: A 1Z0-1122 Guide
The short answer
The OCI AI Foundations certification is Oracle's entry-level AI credential, Oracle Cloud Infrastructure AI Foundations Associate, earned by passing exam 1Z0-1122-26, which Oracle MyLearn listed as the current code on 11 October 2026. It tests concepts, not coding: AI, machine learning and deep learning basics, generative AI and large language models, and which Oracle Cloud Infrastructure (OCI) service fits which job, including OCI Generative AI and Oracle Database AI Vector Search. You prepare with Oracle's own learning path on MyLearn and an independent course, book the exam with Oracle, and pay any exam fee to Oracle. Check the code, question count, time, passing score and fee, if any, on Oracle's exam page on the day you book.
Checked in October 2026. Oracle's education site, education.oracle.com, did not respond when we checked on 11 October 2026, so this guide does not quote the exam fee, question count or pass mark; read them on Oracle's own exam page before you register.
Who should take the OCI AI Foundations exam?
The exam suits people who need to talk sensibly about AI at work rather than build models from scratch:
- Freshers and engineering graduates who want a first AI credential on their CV before IT services or cloud interviews
- IT support, database and application staff in companies that already run Oracle databases or Oracle applications
- Engineers in manufacturing, energy and utilities who are asked to judge AI proposals from vendors
- Sales, presales and consulting staff who need to match a customer's problem to an AI service
It is not a data scientist's credential. If you want to train and deploy models yourself, take it as the first step and then go deeper with Python and machine learning.
What does exam 1Z0-1122-26 cover?
Oracle lists six topic areas for 1Z0-1122-26. Responsible AI runs through all of them.
Topic area (Oracle, read 11 October 2026) | What you must be able to do |
|---|---|
AI foundations | Tell AI, machine learning, deep learning and generative AI apart; name AI workloads |
Machine learning foundations | Supervised, unsupervised and reinforcement learning; read metrics such as precision and recall |
Deep learning foundations | Neurons, layers, CNNs for images, RNNs and LSTMs for sequences, transformers |
Generative AI and LLMs | Tokens, embeddings, prompts, RAG versus fine-tuning |
OCI AI portfolio | OCI Language, Speech, Vision, Document Understanding, OCI Data Science, GPU infrastructure |
OCI Generative AI, AI Vector Search, Select AI | Models, dedicated AI clusters, agents, searching data by meaning |
Oracle's OCI Generative AI documentation describes the service as fully managed, covering models for chat, embeddings and reranking, agents that combine models with tools and retrieval, and governance controls; custom models can be fine-tuned and hosted on dedicated AI clusters. Oracle's AI Vector Search overview explains that it lets you query data by meaning rather than keywords, by storing vector embeddings next to business data.
How the OCI exam compares with other cloud AI fundamentals
Indian recruiters often ask which cloud's AI exam to take. The honest answer is: the one your employer or target employer uses. The concepts overlap heavily.
Credential | Vendor | Choose it when |
|---|---|---|
OCI AI Foundations Associate (1Z0-1122-26) | Oracle | Your company runs Oracle Database, Oracle applications or OCI |
AWS Certified AI Practitioner | Amazon Web Services | Your team builds on AWS |
Azure AI Fundamentals | Microsoft | You work in a Microsoft 365 and Azure shop |
Google Cloud Generative AI Leader | Google Cloud | You advise on generative AI strategy on Google Cloud |
Machine learning metrics, how an LLM works, prompting, RAG and responsible AI are the same on all four. Only the service names change. EDWartens has free preparation courses for the AWS, Azure and Google Cloud exams too, so you can switch without starting again.
Worked example: precision, recall and why accuracy misleads
The machine learning questions on foundation exams are conceptual, but you understand them far better once you have done one calculation by hand. This one comes from the course notes.
Scenario. A spam filter is tested on 1,000 emails. It flags 100 as spam, and 80 of those really are spam. In total, 120 of the 1,000 emails were spam.
- True positives (TP) = 80; false positives (FP) = 100 - 80 = 20
- False negatives (FN) = 120 - 80 = 40
- True negatives (TN) = 1,000 - 80 - 20 - 40 = 860
- Accuracy = (80 + 860) / 1,000 = 0.94
- Precision = 80 / (80 + 20) = 0.80
- Recall = 80 / (80 + 40) = 0.667
- F1 = 2 x 0.80 x 0.667 / (0.80 + 0.667) = 0.727
Accuracy of 94 percent looks excellent, yet recall shows that one spam email in three still reaches the inbox. An exam question that says "most emails are not spam" is pointing you at exactly this trap. When missing a positive is costly (fraud, a faulty part, a disease), recall matters most; when a false alarm is costly, precision does.
A second worked example: will the model fit on the GPU?
The OCI AI portfolio topic includes GPU infrastructure. A quick estimate helps you answer "which infrastructure?" questions:
- An 8-billion-parameter model at 16-bit precision uses 2 bytes per parameter: 8 x 10^9 x 2 = 16 GB for the weights alone
- The same model at 4-bit precision: 8 x 10^9 x 0.5 = 4 GB
- A 70-billion-parameter model at 16-bit: 70 x 10^9 x 2 = 140 GB, more than one 80 GB GPU holds
So the 70B model must be split across at least two such GPUs, or quantised. That is why large models run on clusters, and why OCI offers dedicated AI clusters for fine-tuning and hosting.
How to prepare in five weeks
- Week 1: open Oracle MyLearn, enrol in Oracle's free learning path for the exam, and note the exam code and details shown there.
- Week 2: machine learning and deep learning foundations, with the metric calculations above done by hand.
- Week 3: generative AI, prompts, RAG and fine-tuning. Learn the rule: prompt first, RAG for knowledge that changes, fine-tune for behaviour and style.
- Week 4: the OCI services. For each, write one line: what goes in, what comes out, one use case.
- Week 5: take Oracle's practice exam on MyLearn once, map every wrong answer back to a topic, revise, then book.
An OCI Free Tier account helps with the optional console exercises. Oracle's Free Tier documentation describes Always Free resources that do not expire and a trial with credits valid for up to 30 days. OCI Generative AI is offered only in some regions and is billed by use after the trial, so delete anything you do not need.
Avoid any site that offers "real 1Z0-1122 questions". Using them breaks Oracle's candidate rules and can cost you the credential, and the questions are often wrong. Practise on original questions and Oracle's own practice exam.
A free course for the 1Z0-1122-26 exam
The free Oracle Cloud AI Foundations Exam Prep (1Z0-1122) course maps all six topic areas in thirteen study modules, with lessons from IBM Technology, StatQuest, 3Blue1Brown, Cloud Alchemy Academy and Oracle's own channels, original notes, worked problems like the two above, and original practice questions. It is a free course with a verifiable certificate of completion; anyone can check a certificate on our verification page. It is not the Oracle credential, which only Oracle awards after its exam, and EDWartens is not an Oracle training partner.
For background, read machine learning vs deep learning and what engineers should know about generative AI and LLMs. Then start the OCI AI Foundations exam prep course.
Frequently asked questions
Q: What is the OCI AI Foundations certification?
A: It is Oracle Cloud Infrastructure AI Foundations Associate, Oracle's entry-level AI credential. You earn it by passing Oracle's exam, listed as 1Z0-1122-26 on Oracle MyLearn on 11 October 2026. It tests AI, machine learning, deep learning, generative AI and the OCI AI services at concept level.
Q: Is the OCI AI Foundations exam free?
A: Do not assume so. Any exam fee is set by Oracle and paid to Oracle, and offers can change. Check the current fee, if any, on Oracle's exam page before you register.
Q: What is the exam code for OCI AI Foundations in 2026?
A: 1Z0-1122-26, which replaced 1Z0-1122-25. Oracle refreshes the exam roughly yearly and changes the suffix, so read the code on Oracle's exam page before you book.
Q: Do I need coding to pass 1Z0-1122?
A: No. The exam is multiple choice and conceptual. Basic school algebra helps with metrics such as precision and recall, but you will not write code.
Q: How long does it take to prepare for OCI AI Foundations?
A: About four to six weeks at a few hours a week for a beginner. EDWartens' free prep course takes about 14 hours, and Oracle's own learning path and practice exam add a few more.
Q: Is there a free OCI AI Foundations course with a certificate?
A: Yes. EDWartens' Oracle Cloud AI Foundations Exam Prep (1Z0-1122) is a free course with a verifiable certificate of completion. It prepares you for Oracle's exam; the Oracle credential itself comes only from Oracle.
Learn this, free
The courses that teach this
Every lesson, the written notes and the practice are free with an account. Only the certificate is optional and paid.
