An NVIDIA certification costs $125 for an associate exam and $200 to $500 for a professional exam, paid per attempt when you book in NVIDIA's certification center. Some countries and regions add tax on top. A certification stays valid for two years, and keeping it means passing the exam again at its full price.
Fees and rules below come from NVIDIA's certification page, its FAQ and the individual exam pages, checked October 21, 2026.
What each NVIDIA exam costs
| Exam | Area | Level | Exam fee | Exam time |
|---|---|---|---|---|
| NCA-AIIO: AI Infrastructure and Operations | AI infrastructure | Associate | $125 | 1 hour |
| NCP-AII: AI Infrastructure | AI infrastructure | Professional | $400 | 2 hours |
| NCP-AIO: AI Operations | AI infrastructure | Professional | $500 | 2 hours |
| NCP-AIN: AI Networking | AI infrastructure | Professional | $400 ($500 from November 1, 2026) | 2 hours |
| NCP-ARI: AI Rack and Interconnect | AI infrastructure | Professional | $400 | 2 hours |
| NCA-ADS: Accelerated Data Science | Data science | Associate | $125 | 1 hour |
| NCP-ADS: Accelerated Data Science | Data science | Professional | $200 | 2 hours |
| NCA-GENL: Generative AI LLMs | Generative AI | Associate | $125 | 1 hour |
| NCA-GENM: Generative AI Multimodal | Generative AI | Associate | $125 | 1 hour |
| NCA-AMAD: AI Models and Application Development | Generative AI | Associate | $125 | 1 hour |
| NCP-GENL: Generative AI LLMs | Generative AI | Professional | $200 | 2 hours |
| NCP-AAI: Agentic AI | Generative AI | Professional | $200 | 2 hours |
| NCP-OUSD: OpenUSD Development | Simulation and physical AI | Professional | $200 | 2 hours |
Every associate exam costs $125. The professional exams for software work (data science, generative AI, agentic AI and OpenUSD) cost $200, the infrastructure professional exams cost $400, and NCP-AIO costs $500, with three hands-on lab exercises alongside 30 multiple-choice questions in 120 minutes. NCP-AIN follows on November 1, 2026: NVIDIA says it adds lab exercises to that exam on that day, and the price rises from $400 to $500.
The NVIDIA certifications guide explains what each exam covers and who it is for.
Try 15 NCP-AAI practice questions free
They come from the NCP-AAI practice tests, and each option shows its explanation once you answer.
- 7 timed practice exams
- Every answer explained
- A score for each domain
- $19.99 one-time
NVIDIA generative AI certification cost
Five of the thirteen exams cover generative AI. The three associate exams, NCA-GENL for large language models, NCA-GENM for multimodal models and NCA-AMAD for building AI applications, cost $125 each. The two professional exams, NCP-GENL for optimizing and deploying LLMs and NCP-AAI for agentic AI, cost $200 each. The generative AI certification comparison helps you pick between them.
What a retake, a cancellation or a renewal costs
- Retakes. If you fail, you buy the exam again at the full fee. You wait 14 days between attempts, and you can take the same exam at most five times in 12 months, counted from the day you bought the first attempt.
- Cancellations. You can cancel up to 24 hours before the session. Inside 24 hours the exam can't be cancelled or refunded, and NVIDIA doesn't refund a missed exam or one you have taken, pass or fail.
- Rescheduling. You can move the exam if you do it at least 24 hours before the session, and the new date has to fall within two months of the day you paid.
- Renewal. A certification is valid for two years from issue. To recertify, you retake the exam, so plan for the exam fee again every two years.
- Tax. Fees are listed in US dollars, and NVIDIA says additional taxes may apply in some countries and regions.
Three questions from the NCP-AAI practice tests
Your team must pick an agent framework for three products: a retrieval-heavy document assistant, a role-based research crew of four agents and a payments workflow with strict, custom approval steps. Last year's prototype took two weeks to debug a looping agent, because its framework exposed no traces. Select TWO criteria that should drive the framework decision. (Select TWO)
[object Object]
Correct. Debuggability dominates production operations, so tracing that exports to OpenTelemetry and works with NeMo Agent Toolkit, Phoenix, Weave or Langfuse would have exposed the looping agent quickly.
Correct. A framework whose primitives match the workload stays a thin layer: retrieval-oriented libraries such as LlamaIndex for documents, role-based crews such as CrewAI for research and custom code where stock primitives strain, as in payments.
A. A fully custom stack means rebuilding tracing, retries, tool routing and eval harnesses yourself; custom code fits where frameworks strain.
C. Switching frameworks often means refactoring consumers, retraining developers and redoing evals each time.
D. Popularity says little about fit for a given workload or about production observability.
Your order pipeline has five agents (intake, fraud check, inventory, shipping, notifications) deployed as separate services that scale independently. Today intake calls each downstream agent over synchronous HTTP; when the shipping agent restarts, orders fail and intake retries pile up. New agents, such as a loyalty-points agent, will need the same order events. Which communication approach best fits this system?
[object Object]
Correct. A broker decouples producers from consumers: it holds messages while the shipping agent restarts, fans each event out to every subscriber and lets a new loyalty agent subscribe without changes to intake.
B. Retries soften brief failures, but intake still waits on each agent and must change for each new consumer.
C. In-process calls avoid network failures but give up independent scaling, and one exception can stop the whole pipeline.
D. A file on one host is a single point of failure, with no delivery guarantees or access from other hosts.
Your long-running assistant's conversations exceed the context window within days. Before old turns are evicted, the team wants to keep what matters: in one week a user gave their address, retracted an email address, chose H100 nodes for a project and chatted about the time of day. Select TWO design choices that produce useful memory consolidation. (Select TWO)
[object Object]
Correct. Structured facts such as (user, ships_to, 22 Maple Ave, chat, date) give long-term memory a queryable schema that can be deduplicated and updated, separate from conversational text.
Correct. Keeping the address, the H100 decision and task outcomes while dropping small talk and the retracted email keeps the store accurate and useful for retrieval.
A. Durability differs by type: an address should last and small talk should not, so one expiry fits neither.
D. Model weights do not learn from conversations at run time, so anything not stored is lost.
E. A verbatim archive keeps chatter and retracted statements that pollute retrieval.
Full NCP-AAI set: 7 timed exams, every answer explained, $19.99 one-time.
See the practice testsDiscounts on NVIDIA exams
NVIDIA offers educational and group discounts, which you request through the contact link in the FAQ on its certification page. NVIDIA is also holding a certification webinar on November 19, 2026, in two sessions (9 a.m. CET and 9 a.m. PT), and attendees receive a 50% off exam discount code. On a $200 professional exam, that saves $100.
What preparation costs
NVIDIA lists recommended training on each exam page. The associate pages for NCA-GENL and NCA-GENM start their learning paths with NVIDIA's Fundamentals of Deep Learning, published as free open-source materials, and they list suggested reading for every exam topic, most of it free NVIDIA documentation and blog posts.
The two professional generative AI exams print their course prices on the exam page:
| Course | Format | Price | Length | On the path for |
|---|---|---|---|---|
| Building RAG Agents With LLMs | Self-paced | $90 | 8 hours | NCP-AAI, NCP-GENL |
| Building Agentic AI Applications With LLMs | Self-paced | $90 | 8 hours | NCP-AAI |
| Introduction to Deploying RAG Pipelines for Production at Scale | Self-paced | $90 | 8 hours | NCP-AAI |
| Evaluating RAG and Semantic Search Systems | Self-paced | $30 | 3 hours | NCP-AAI |
| Adding New Knowledge to LLMs | Instructor-led | $500 | 8 hours | NCP-AAI, NCP-GENL |
| Model Parallelism: Building and Deploying Large Neural Networks | Instructor-led | $500 | 8 hours | NCP-GENL |
| Deploying RAG Pipelines for Production at Scale | Self-paced | $500 | 8 hours | NCP-GENL |
| Optimizing CUDA ML Codes With NVIDIA Nsight's Profiling Tools | Self-paced | $30 | 4 hours | NCP-GENL |
Taking every course on the NCP-AAI path costs $800, and every course on the NCP-GENL path costs $1,620. The NCP-GENL page lists its RAG deployment course as self-paced at $500, while the NCP-AAI page lists the introductory version at $90, so check the course page before you buy. If you already do the work an exam covers, skipping the courses on topics you know is the easiest saving.
The infrastructure exams link to their courses instead of printing prices. NCA-AIIO's path is the self-paced AI Infrastructure and Operations Fundamentals course, about seven hours long. NCP-AII and NCP-AIO add a multi-day professional workshop to that course, and NCP-AIN lists three self-paced courses on InfiniBand and Cumulus Linux. NVIDIA shows those prices on each course page.
Practice tests are another optional cost. preporato.com has practice tests for every exam in the table, at $19.99 one-time per exam or as part of Pro, and each has free sample questions you can take first.
Three example budgets
| Plan | Exam | Training | Total |
|---|---|---|---|
| NCA-GENL with NVIDIA's free materials | $125 | $0 | $125 |
| NCP-AAI with the four self-paced courses | $200 | $300 | $500 |
| NCP-AAI with its full learning path | $200 | $800 | $1,000 |
Each failed attempt adds the exam fee again, and staying certified adds it once more every two years. For the free side of the budget, see what you can get from NVIDIA for free, and for one exam in detail, the NCP-AAI cost and ROI breakdown.
Frequently asked questions
Sources:
- NVIDIA certification program and FAQ
- NVIDIA-Certified Professional: Agentic AI
- NVIDIA-Certified Professional: Generative AI LLMs
- NVIDIA-Certified Associate: Generative AI LLMs
- NVIDIA-Certified Associate: AI Infrastructure and Operations
- NVIDIA-Certified Professional: AI Operations
- NVIDIA-Certified Professional: AI Networking