Why Use GitHub Copilot?
GitHub Copilot is known for its auto-fill technology. The World Wide Web is filled with articles and videos claiming that it is only useful for code completion and lacks usefulness as an agentic coding tool.
My experience is that GitHub Copilot is an excellent agentic coding tool. I have used it to build some very large and complex applications in a short period of time. I combine Copilot with a tool that makes it very effective. The best part of Copilot is the price.
Cost
A GitHub Copilot Pro+ membership costs $40/month. If you pay for a yearly membership, the monthly price is around $32.
The plan includes 1500 premium requests per month. A premium request is a user-submitted message to a premium model. Several API calls may result from a single premium request.
I consider myself to be a heavy user. I average around 700 premium requests per month. The 1500 request limit is more than enough. This is especially true if you use the best model for the job.
Models
I use the following models in my workflow:
- Claude Sonnet 4.5: This is my primary model. It is great for general-purpose backend work.
- Claude Opus 4.5: This is my planner. It costs more credits, so it should be used sparingly.
- Gemini 3 Pro: This is my web designer. It is good at creative tasks.
- GPT 5.2 Codex: This is my code reviewer and bug fixer.
Smaller Context Windows
One reason Copilot is so affordable compared to other products on the market is that its models use smaller context windows. This is problematic for traditional vibe coding. However, using the proper tools will resolve the limitation.
Introducing OpenSpec
To use Copilot effectively, you must introduce a planning tool. OpenSpec is more than just a planning tool. In addition to creating plans, known as change proposals, it creates project specifications. This will aid the model in future planning sessions.
To start, you must install the OpenSpec CLI:
npm install -g @fission-ai/openspec@latest
Next, you must activate OpenSpec in each project. Navigate to your project directory and run:
openspec init
Press ENTER next to GitHub Copilot to enable it, and press TAB to proceed. The OpenSpec CLI will generate multiple files and directories that work together to create the OpenSpec experience.
IMPORTANT: After configuring OpenSpec for a project, you must restart Visual Studio Code so it will read the OpenSpec files.
Your First Change Proposal
Now that OpenSpec is initialized, open Copilot Chat, select Claude Opus 4.5, and type:
/opsx-ff my-new-feature Create a change proposal and its artifacts. Add the ability for the user to sort rows by most recent or by chronological order.
OpenSpec will proceed to generate the change proposal in the openspec directory. Open the tasks.md file and look at the tasks OpenSpec produced. If everything looks good, proceed to apply the proposal. Start a new chat session, select the most appropriate model, and run this command:
/opsx-apply my-new-feature
This will complete the tasks and mark them as completed.
Conclusion
Using OpenSpec, GitHub Copilot generated a large feature with only a few premium requests. This reduced the number of requests needed to implement the solution. It also improved the accuracy of the results. Using OpenSpec with GitHub Copilot offers one of the most affordable ways to access premium large language models.
