We earn a commission if you buy through links on this page. How we're funded.
Quick answer
Oracle's OCI Generative AI Professional certification is better than most readers expect and narrower than Oracle implies, and it may not be offered much longer: when we checked on 26 September 2026, Oracle's certification list no longer showed it, its Japanese exam was due to end on 30 September 2026, and its 2025 learning path was to be archived that day. Oracle now lists the OCI Enterprise AI Professional (1Z0-1158-26), a newer exam weighted toward agents: 90 minutes, 50 questions, 68% to pass. Take either if your employer runs Oracle Cloud; outside Oracle shops, the signal is thinner than the big-three cloud credentials.
AI-102 is retired. The replacement is Microsoft Certified: Azure AI Apps and Agents Developer Associate (AI-103).
The table below compares 5 certifications on provider, level, realistic time, coding needed and best for. The first row is a course from our affiliate partners that we chose for this page.
| Certification | Provider | Level | Realistic time | Coding needed | Best for | Enrol |
|---|---|---|---|---|---|---|
| Generative AI with Large Language Models | DeepLearning.AI & AWS | Intermediate | ~17 hours | Yes (Python) | The LLM lifecycle and fine-tuning, outside Oracle | Coursera → |
| OCI Generative AI Professional | Oracle | Professional | ~4–6 weeks part-time for developers | Yes (Python helpful) | The certification this review covers | |
| IBM Generative AI Engineering Professional Certificate | IBM (Coursera) | Intermediate | ~188 hrs | Yes (Python) | The vendor-neutral generative-AI alternative | Coursera → |
| AWS Certified AI Practitioner (AIF-C01) | AWS | Foundational | ~4–6 weeks of prep | No | The AWS-side credential if your cloud is elsewhere | |
| Azure AI Apps and Agents Developer Associate (AI-103) | Microsoft | Intermediate (associate) | ~1–2 months of prep | Yes (Python) | The Azure-side practitioner counterpart (it replaced AI-102) |
What is the OCI Generative AI Professional certification?
It is Oracle's professional-level credential for building generative-AI applications on Oracle Cloud Infrastructure. Unlike the fundamentals-tier certificates that dominate this space, it targets working developers: the published scope covers LLM concepts and the practical use of OCI's Generative AI service — prompting, fine-tuning options, and retrieval-augmented generation patterns.
Two things distinguish it from the crowd. First, it went deep on generative AI earlier than most vendor certifications, so the syllabus reads current rather than retrofitted. Second, Oracle has used it as a marketing instrument for OCI — which is exactly why the free windows discussed below exist, and why you should read the signal it sends with open eyes.
Is it being retired? What Oracle lists now
The signs point that way, though Oracle has not said so for the English exam. When we checked on 26 September 2026, Oracle's certification list no longer showed the OCI Generative AI Professional; the only version of its exam we could find, the Japanese 1Z0-1127-25, said it would end on 30 September 2026; and Oracle's 2025 learning path for it said it “will be archived on 30-Sep-2026”. The English exam page could not be reached that day, because Oracle's education site was down for maintenance.
What Oracle's list does show is a newer exam, labelled NEW: Oracle Cloud Infrastructure Enterprise AI Professional (1Z0-1158-26). Oracle does not say it replaces the Generative AI Professional, and it is a different exam. Its topics table puts most of the weight on agents — applying agentic AI concepts (20%), building OCI Enterprise AI agents (25%), and building and deploying hosted agents (15%) — beside large language model concepts (10%), configuring OCI Enterprise AI models (20%) and governing and optimising them (10%). Oracle's official preparation is a learning path of nine-plus hours, and no Oracle page we read says it is free.
The table below shows the fee and format of the Oracle Cloud Infrastructure Enterprise AI Professional exam, as Oracle publishes them. Every other exam we track is compared on our AI certification exams page.
| Credential | Vendor | Level | Fee | Exam format | Enrol |
|---|---|---|---|---|---|
| Oracle Cloud Infrastructure Enterprise AI Professional Labelled NEW by Oracle. | Oracle | Professional | Shown only after you sign in to register | 90 minutes · 50 questions · 68% to pass · online proctored | Oracle → |
Generative AI Professional vs Enterprise AI Professional
If you came here choosing between these two Oracle professional exams, this is the comparison in one place. The Generative AI Professional column is what Oracle showed on 26 September 2026, and the Enterprise AI Professional column comes from its MyLearn exam page the same day. Oracle does not say the second replaces the first.
The table below compares the OCI Generative AI Professional and the OCI Enterprise AI Professional on whether Oracle still lists each, what each centres on, what Oracle publishes about the exam, whether Oracle's free programme covered it, and who should take it.
| Question | OCI Generative AI Professional (1Z0-1127-25) | OCI Enterprise AI Professional (1Z0-1158-26) |
|---|---|---|
| Does Oracle's certification list show it? | No, when we checked on 26 September 2026; its Japanese exam was due to end on 30 September 2026 | Yes, labelled NEW |
| What does it centre on? | Large language model concepts and Oracle's OCI Generative AI service: prompting, fine-tuning options and retrieval-augmented generation | Agents: applying agentic AI concepts and building and deploying agents carry most of the weight in its topics table |
| What does Oracle publish about the exam? | Its English exam page could not be reached on 26 September 2026, so we publish no format | Duration, question count and pass mark, in the Oracle table above; the fee only after you sign in |
| Was it in Oracle University's free programme on 26 September 2026? | No | No |
| Who should take it? | A developer in an Oracle shop building LLM applications, if Oracle still lets them book it in their language | A developer in an Oracle shop building agents on OCI, or anyone who finds the Generative AI Professional no longer offered |
The practical rule: check first whether Oracle still lets you book the Generative AI Professional in your language. If it does and your work is LLM applications on OCI, it is the closer fit to that work; if it does not, or your team is building agents, the Enterprise AI Professional is the exam Oracle lists now. Both certifications last 24 months, like every Oracle Cloud Infrastructure certification, and outside Oracle shops neither substitutes for your own cloud's credential.
What does the exam actually cover?
Study from Oracle's official exam page, not third-party summaries. The published scope spans LLM fundamentals — architectures, prompting, common failure modes — and the applied layer: using the OCI Generative AI service, building RAG pipelines against OCI data stores, and the security and cost considerations of running generative workloads. The same exam page gives the current question count, duration and passing score, and those are the figures to plan around.
The honest content assessment: the LLM material is solid and transfers anywhere; the service-specific material transfers only if you work on OCI. That ratio — portable theory to ecosystem plumbing — is roughly the same as AWS and Azure equivalents, so the deciding factor is which plumbing your employer runs.
Who is it for?
Developers and architects in Oracle shops, first and foremost — organisations running Oracle databases, E-Business Suite or OCI infrastructure, where the natural genAI deployment path is Oracle's own service. Data engineers in those environments get direct value: the RAG and pipeline content maps onto work they will actually be asked to do. It also fits Oracle partners and consultancies, where the credential is table stakes for practice credibility.
If you write software outside the Oracle ecosystem, our software engineers guide will route you better — this certification would teach you real things, but on a platform you may never touch professionally.
How is it recognised outside Oracle shops?
Thinly — and pretending otherwise would not serve you. Recruiters pattern-match the credentials they see most, which means AWS, Azure and Google Cloud (the trade-offs between those are in our cloud comparison). An OCI certification on a CV aimed at non-Oracle employers reads as evidence of general genAI literacy, not as a platform qualification they can use.
Inside the Oracle world the picture inverts: Oracle-heavy enterprises, Oracle partners and consultancies read it immediately, and it can be the differentiating line. Signal strength is a function of who is reading — this credential has one of the sharpest inside/outside splits in the AI certification market.
How hard is it, and how long does preparation take?
Moderate for a working developer: harder than the fundamentals tier (no-code, concept-only exams), gentler than the cloud professional tier. Four to six weeks of part-time preparation is realistic if you can already write some Python and have touched a cloud console; add time if RAG and embeddings are new vocabulary. Check the passing score and retake policy on Oracle's exam page before you book.
The efficient prep path mirrors our advice for every practitioner exam: work through Oracle University's official course for the certification, then build one small RAG application against the OCI free tier before you sit the exam. The scenario questions are far easier when you have made the implementation mistakes once already.
What does it cost — and the free-window question
Outside a promotion you pay Oracle's exam fee. Oracle's exam pages show it only after you sign in to register, priced for your country, so we publish no figure; the official training is priced separately. But the number to actually check is the promotion calendar: Oracle has periodically opened free-certification windows — training and exam at no cost — as part of its push to grow the OCI ecosystem. The free programme Oracle University listed when we checked on 26 September 2026, “Free Training and Certifications from Oracle University”, covered Foundations and Associate certifications, among them OCI AI Foundations Associate and Oracle Agentic AI Foundations Associate, and included neither the Generative AI Professional nor the Enterprise AI Professional.
Those windows change the calculus completely. At zero cost, this certification is one of the best free deep-dives into applied generative AI from any major vendor, worth taking for the knowledge alone regardless of your stack. At full price, it needs the Oracle-shop justification.
Not sure this is the right one for you?
Tell the picker about your background and what you want the certificate to do, and it narrows the list to the one or two courses we would start with. It suggests only our affiliate partners’ courses, and says so before it suggests anything.
Try the AI Certification Picker →What should non-Oracle readers take instead?
Match the substitute to what attracted you here. If it was the generative-AI depth, IBM's Generative AI Engineering certificate covers RAG, embeddings and LLM app patterns vendor-neutrally, and our generative AI certification roundup ranks the wider field. If it was the idea of a platform credential, take the one matching your employer's actual cloud — the comparison guide above will settle which. There is no version of this decision where certifying a platform you will never touch beats certifying the one you use daily.
Our take: buy the knowledge, price the signal
Most coverage of this certification splits into Oracle marketing and reflexive dismissal, and both miss it. The dismissal — 'nobody uses OCI' — undersells a genuinely well-built genAI syllabus and one of the most generous free-training instruments any vendor has run. The marketing oversells the career effect: a certification's signal is only as strong as the hiring market that reads it, and outside the Oracle ecosystem that market is small.
So our position is unusually simple. When a free window is open, the knowledge-to-cost ratio is unbeatable — take it, learn the portable LLM material, and list it as what it is. At full price, take it when Oracle pays your bills or plausibly will. And never take it as a substitute for the credential your own stack's employers actually search for.
Verdict
For developers in Oracle shops: the OCI Generative AI Professional's syllabus maps directly onto the genAI work your organisation will deploy, and the credential reads well where you already are — but check first that Oracle still offers it in your language, and if it does not, look at the Enterprise AI Professional, which Oracle now lists and which weighs agents most. For everyone else: take one only during a free window, for the knowledge; the durable signal for your CV lives on the cloud your employers actually run, and the staged path in our AI certification roadmap — or a two-minute run through the Picker — will name it. Our 2026 ranking shows where Oracle's credential sits in the wider field, and our analysis of whether AI certifications are worth it explains the ecosystem-signal effect in full.
Certifications featured in this guide
Every option below is one we cover in depth. Each link goes to the provider’s own page; where we’ve published a full review, read that first.
Frequently asked questions
Is the OCI Generative AI Professional being retired?
Probably, though Oracle has not said so for the English exam. When we checked on 26 September 2026, Oracle’s certification list no longer showed the OCI Generative AI Professional, its Japanese exam (1Z0-1127-25-JPN) said it would end on 30 September 2026, and Oracle’s 2025 learning path for it said it “will be archived on 30-Sep-2026”. The English exam page could not be reached that day, because Oracle’s education site was down for maintenance.
Oracle now lists the OCI Enterprise AI Professional (1Z0-1158-26), labelled NEW: 90 minutes, 50 questions, 68% to pass, proctored online. Oracle does not call it a replacement, and it is a different exam, weighted toward building and running agents, so check its topics against what you studied before you book it.
Is the Oracle OCI Generative AI certification worth it?
In an Oracle-stack job or partner practice, yes — the syllabus matches real deployment work and the credential reads well internally. Outside the Oracle ecosystem, it is worth taking mainly during free-certification windows for the genuinely strong LLM and RAG content, not for CV signal.
Separating the content from the signal is the honest way to judge this one. The material on retrieval, embeddings and deployment is solid and would be worth studying wherever it came from; the credential is recognised by Oracle shops and largely invisible elsewhere. If you are inside that ecosystem you get both halves. If you are not, you are taking a good course that happens to end in an exam, which is fine so long as you know that is the purchase.
Is the OCI Generative AI exam ever free?
Oracle has run free-certification promotions covering both training and the exam, but not for this exam right now. The free programme Oracle University listed when we checked on 26 September 2026, “Free Training and Certifications from Oracle University”, covered Foundations and Associate certifications, including OCI AI Foundations Associate and Oracle Agentic AI Foundations Associate, and neither the Generative AI Professional nor the newer Enterprise AI Professional. Check Oracle University's current offers before paying.
Those windows are the reason this credential appears on shortlists it would not otherwise reach, and they are worth watching for rather than waiting on. They are announced with limited notice and run for a set period, so the practical approach is to know roughly what you would study and be ready to move when one opens — not to defer a credential you actually need on the chance that a promotion arrives.
How hard is the OCI Generative AI Professional exam?
Moderate: harder than no-code fundamentals exams like the AWS AI Practitioner, gentler than professional-tier cloud ML exams. A developer with some Python and cloud familiarity typically needs four to six weeks part-time. Candidates without RAG or embeddings exposure should budget extra time for the applied material.
The applied half is where the time actually goes, and it is the half worth having. Retrieval questions assume you have some feel for what chunking does to answer quality and why a similarity search returns the wrong passage — intuitions built by constructing something rather than by reading. Build a small retrieval application before the last fortnight of preparation and the applied questions become much easier.
Does the Oracle AI certification expire?
Yes. Oracle’s certification guidelines say Oracle Cloud Infrastructure certifications are valid for 24 months from the date the credential is earned, and its exam pages point to its cloud recertification policy for keeping one current. Factor the renewal into the value calculation before you book.
Factor it in properly if you are taking this during a free window, because the promotion covers the first sitting and not the renewals that follow. A credential that is free once and chargeable thereafter is still worth having — but the honest plan is either to keep renewing it because you work in the ecosystem, or to treat it as a dated line that will eventually lapse, rather than to be surprised later.
What is the best alternative to Oracle's generative AI certification?
Vendor-neutral: IBM's Generative AI Engineering Professional Certificate, which covers RAG, embeddings and LLM application patterns without tying you to one cloud. Platform-matched: the AI credential of whichever cloud your employer runs — AWS, Azure or Google Cloud — since ecosystem-bound signals pay best on the ecosystem you actually use.
Choose between those two on whether your next job is likely to be at your current employer. If it is, the platform credential wins outright because it maps to the work in front of you. If you expect to move, the vendor-neutral one travels — and the concepts underneath are the same in every case, which is why the choice is about signalling rather than about what you will learn.
Keeping this current. Course formats, prices, and certification exam fees change and vary by region. We review our guides regularly, and we always recommend confirming the specifics on the provider's official page before you enrol.