Google Cloud Professional Agentic Architect

It has two parts. First you take a proctored multiple-choice exam delivered by Pearson, online or at a test center. If you pass, you get access to a separate set of hands-on labs in Google Skills and have two months to complete them. You earn the credential only after passing both parts.

The generally available exam follows Google's standard professional format: a 2-hour session, with official results within 7 to 10 business days. Google has not published the question count or passing score for it. The beta, which ran in September 2026, had about 80 multiple-choice questions in 3 hours.

The generally available exam costs 200 US dollars plus tax where applicable. The beta was offered at 120 US dollars.

Google says agent technology changes too fast for the usual two-year term of its professional certifications. It describes a continuous education path to keep the certification active.

Google's exam guide lists the Agent Development Kit, Agent Runtime (formerly Agent Engine), Agent Search (formerly Vertex AI Search), RAG Engine, Vector Search and Agent Retrieval, Agent Identity, Agent Registry, Agent Gateway, Model Armor, the Agents CLI, Antigravity, Gemini Enterprise and core services such as Cloud Run, GKE, BigQuery and Cloud Trace. Vertex AI itself is now called Gemini Enterprise Agent Platform, so older study material often uses outdated names.

Yes. The exam guide names Claude Code on Google Cloud next to Antigravity in the section on using coding agents, which covers configuring them with MCP servers, skills and tools and running them in secure sandboxes.

No. Google offers a learning path on Google Skills as preparation, but it is not required. The certification labs you take after passing the exam are a separate set from the labs in the learning path.