GH-600 Exam Guide: GitHub Certified Agentic AI Developer
Everything on the GH-600 in one place — the exam format, the six skills-measured domains and how much each one counts, and what each domain actually asks of you. Written in plain English from the official objectives, so you know exactly what to study.
This page is the map. The lessons are the journey — free, beginner-friendly, one idea at a time. Start the free GH-600 course →
What is the GH-600 exam?
The GH-600 (GitHub Certified: Agentic AI Developer) is GitHub's certification for people who build, run, and supervise AI agents inside real software projects — using GitHub as the place where the work is tracked and controlled.
In plain terms, an "agent" here is an AI helper (think GitHub Copilot and similar tools) that can take on a task, plan it, do work like writing code or opening a pull request, and check its own results. The exam tests whether you can put these agents to work safely and usefully: giving them the right tools, keeping them inside guardrails, watching what they do, and fixing them when they go wrong.
Officially, the audience is someone comfortable with the software development lifecycle (how code goes from idea to shipped), with GitHub's workflows and controls, and with coding agents such as Copilot, MCP servers, and agent customization. If that sounds like a lot of unfamiliar words — that's exactly what the free course is for. It starts from the very beginning.
GH-600 exam format at a glance
| Exam name | GitHub Certified: Agentic AI Developer |
| Exam code | GH-600 |
| Passing score | 700 out of 1000 |
| Length / questions | Timed; commonly reported around 120 minutes and ~40–60 questions (confirm on the official exam page) |
| Status | Beta — the credential you earn is the same as the general-availability exam |
| Cost to study | Free, here at agenticlately |
| Languages | If the exam isn't offered in your language, you can request an extra 30 minutes |
The passing score (700/1000) and the domain weights below come straight from the official Microsoft Learn study guide. Exam length and question count aren't published there, so treat those as widely-reported estimates and confirm on the official page before you book.
The six GH-600 domains (and how much each counts)
The exam is split into six skills areas. The percentages are the share of the exam each one is worth, so they tell you where to spend your study time. Domain 2 (tool use) is the heaviest.
| # | Domain | Weight |
|---|---|---|
| 1 | Prepare agent architecture and SDLC processes | 15–20% |
| 2 | Implement tool use and environment interaction | 20–25% |
| 3 | Manage memory, state, and execution | 10–15% |
| 4 | Perform evaluation, error analysis, and tuning | 15–20% |
| 5 | Orchestrate multi-agent coordination | 15–20% |
| 6 | Implement guardrails and accountability | 10–15% |
1. Prepare agent architecture and SDLC processes (15–20%)
Fitting agents into how software actually gets built: deciding what steps an agent should handle, keeping its planning separate from its doing (and checking the plan before it acts), and setting how much freedom it gets so a human can step in without slowing everything down.
2. Implement tool use and environment interaction (20–25%)
The biggest domain. Choosing and setting up the tools an agent can use, connecting tools through MCP servers (a standard way to plug capabilities into an agent), scoping an agent to the right repository or branch, and handling things safely when they fail — retries, rollbacks, and a clear trail of what the agent did.
3. Manage memory, state, and execution (10–15%)
How an agent remembers. Choosing what it should keep for the short term, the long term, or store elsewhere; saving progress so it can pick up where it left off without redoing work; and stopping it from drifting or acting on stale information.
4. Perform evaluation, error analysis, and tuning (15–20%)
Measuring whether an agent did a good job. Setting clear success criteria, reading logs and traces to find why something went wrong, and then adjusting the instructions, memory, or tools to make it better next time.
5. Orchestrate multi-agent coordination (15–20%)
Running more than one agent together. Coordinating them, keeping them out of each other's way, spotting and resolving conflicts (like two agents editing the same code), and adding, replacing, or retiring agents without breaking work that's already running.
6. Implement guardrails and accountability (10–15%)
Keeping agents safe and answerable. Deciding which actions truly need a human's approval, blocking anything that breaks security or compliance rules, giving each agent only the access it needs, and requiring sign-off for risky or hard-to-undo changes.
How to study for the GH-600
The fastest honest path, especially if you're new to this:
- Skim the official objectives so you know the shape of the exam — the Microsoft Learn study guide.
- Learn it from zero with the free lessons. If you're newer, begin with GitHub basics (Phase 0), then work through the six domains in order.
- Quiz yourself — every lesson ends with a short quiz, and the site tracks what you've got down.
- Trust what you learn — every fact links back to its official source, so you're never memorizing something that turns out to be wrong.
Open the free GH-600 course → No signup, no paywall.
Official GitHub & Microsoft training
Microsoft Learn also offers free self-paced modules for the GH-600. Use them alongside the lessons here:
GH-600 FAQ
Is the GH-600 exam free?
The exam itself is a paid GitHub certification, but you can study for it completely free with the course here.
What is the GH-600 passing score?
700 out of 1000, per the official Microsoft Learn study guide.
How hard is the GH-600?
It covers six areas, from agent architecture to guardrails. With structured prep it's very approachable, even with no AI background — the lessons here are built for total beginners and define every term as it comes up.
Do I need an AI or coding-certification background?
No. The course assumes no prior AI or certification experience and starts with GitHub basics.
Is this an official study guide?
No — see the note below. It's an independent, free resource. The official guide lives on Microsoft Learn.
More GH-600 guides
- GH-600 practice questions — free scenario questions with explained answers.
- Is the GH-600 worth it? — who it's for, what it costs, and how hard it is.
- The full free GH-600 course → — learn it from zero, one lesson at a time.
Unofficial. Not affiliated with, authorized, endorsed by, or sponsored by GitHub, Inc. or Microsoft Corporation. “GitHub”, “GitHub Copilot”, and “GH-600” are trademarks of their respective owners, referenced here for identification and study purposes only. Exam details can change — always confirm against the official Microsoft Learn study guide.