What changed between MLA-C01 and MLA-C02? +
AWS kept the same domain structure but rewrote the task statements and skills inside each domain. C02 adds fine-tuning foundation models, RAG architectures and agent orchestration, and expands Amazon Bedrock coverage alongside Amazon SageMaker AI. The last day to take MLA-C01 in English was September 28, 2026. Courses written for C01 leave the new generative AI scope largely uncovered.
What is the MLA-C02 beta exam format? +
The C02 beta has 85 questions in 170 minutes, multiple choice and multiple response, in English only. The extra length comes from additional unscored items used to set the final exam. It is delivered at Pearson VUE centers or online. Beta delivery began September 29, 2026; AWS has not yet announced the standard (GA) release dates.
How much does MLA-C02 cost? +
During the beta, the exam costs USD 75 - half the USD 150 standard price of MLA-C01. The certification is valid for three years either way. AWS says beta results are typically available within five business days, and it has not yet published the GA date. Check the AWS certification page for the standard price and scoring details once C02 reaches general availability.
Should I take the beta or wait for the standard MLA-C02? +
Take the beta if you are confident on SageMaker AI and Bedrock and want the lower fee. Wait if you prefer a shorter exam with more mature prep material, or need a language other than English - Japanese, Korean and Simplified Chinese MLA-C01 remain available until C02 reaches GA. Beta takers sit 85 questions instead of the usual 65.
How hard is MLA-C02 and what makes it hard? +
The challenge is breadth. You must handle data preparation, model training and tuning, deployment, monitoring and security in SageMaker AI, and now judge when to customize a foundation model in Bedrock instead. Skimmers who compare only domain names assume nothing changed and miss the rewritten skills the new questions test. Working ML engineers typically need 6-10 weeks.
What experience does MLA-C02 assume? +
The exam targets people who build, deploy and operate ML workloads on AWS, roughly a year of hands-on experience with SageMaker AI and related services. With C02, experience calling foundation models through Bedrock, building RAG pipelines and evaluating fine-tuned models also matters. If you have only trained models in notebooks, spend time on deployment, monitoring and CI/CD for ML before booking.