Search...
Everything needed to plan for the Agentic AI Developer (GH-600) (GH-600) exam: number of questions, time limit, passing score, and the official domain weightings. For the study checklist itself, see the GH-600 study guide.
Agentic AI Developer (GH-600) (GH-600) is a Associate level GitHub certification exam with about 60 questions in 120 minutes, requiring 70% to pass. It covers 6 official domains: Implement Tool Use and Environment Interaction (23%); Prepare Agent Architecture and SDLC Processes (17%); Perform Evaluation, Error Analysis, and Tuning (17%); Orchestrate Multi-Agent Coordination (17%); Manage Memory, State, and Execution (13%); Implement Guardrails and Accountability (13%).
Exam code
GH-600
Questions
~60
Time limit
120 min
Passing score
70%
Official domains
6
Level
Associate
Provider
GitHub
Practice questions
300+ on CertStud
| Domain | Weight | What it covers |
|---|---|---|
| Implement Tool Use and Environment Interaction | 23% | Configure tools and permissions, MCP servers, CI integration, error handling, rollbacks |
| Prepare Agent Architecture and SDLC Processes | 17% | Integrate agents into SDLC, define planning/execution boundaries, configure observability |
| Perform Evaluation, Error Analysis, and Tuning | 17% | Define success criteria, analyze failures, classify root causes, tune agent behavior |
| Orchestrate Multi-Agent Coordination | 17% | Orchestration patterns, conflict detection, multi-agent observability, failure recovery |
| Manage Memory, State, and Execution | 13% | Memory types, context drift prevention, state persistence, cross-agent state sharing |
| Implement Guardrails and Accountability | 13% | Autonomy levels, guardrail design, least privilege, human-in-the-loop, audit logging |
Percentages reflect published official domain weightings. Prerequisite knowledge and current scoring rules are confirmed on the GitHub exam page before booking.
The GH-600 exam has approximately 60 questions. Question counts can vary slightly between deliveries, so treat 60 as the planning figure rather than a fixed number.
Candidates get about 120 minutes to complete the GH-600 exam, which works out to roughly 2 minute(s) per question at 60 questions.
The GH-600 exam requires 70% to pass. Scores are reported on a scaled rather than a raw percentage basis, so a passing result does not correspond to a simple count of correct answers.
The GH-600 exam covers 6 official domains: Prepare Agent Architecture and SDLC Processes (17%), Implement Tool Use and Environment Interaction (23%), Manage Memory, State, and Execution (13%), Perform Evaluation, Error Analysis, and Tuning (17%), Orchestrate Multi-Agent Coordination (17%), Implement Guardrails and Accountability (13%).
Implement Tool Use and Environment Interaction carries the heaviest weighting on the GH-600 exam at 23% of the exam. Configure tools and permissions, MCP servers, CI integration, error handling, rollbacks
GH-600 is the exam code for Agentic AI Developer (GH-600) (GitHub Certified: Agentic AI Developer), a Associate level GitHub certification exam covering 6 domains.
The GH-600 exam reports 70% on a scaled score. Plan for roughly 60 questions across 120 minutes, weighted by domain: 23% Implement Tool Use and Environment Interaction, 17% Prepare Agent Architecture and SDLC Processes, 17% Perform Evaluation, Error Analysis, and Tuning, 17% Orchestrate Multi-Agent Coordination, 13% Manage Memory, State, and Execution, 13% Implement Guardrails and Accountability.
CertStud offers free GH-600 practice questions with detailed explanations, plus full-length practice exams and flashcards at https://certstud.com/certifications/github/agentic-ai-developer. Start with 10 free questions at https://certstud.com/try-free.