GitHub Copilot
GitHub Copilot is a general-purpose AI coding assistant that suggests code as you type, answers questions about a codebase, reviews changes and runs multi-step work on its own. It runs inside Visual Studio Code, the same editor an IBM i shop already uses for Code for IBM i, but GitHub trains it on public code and says suggestion quality depends on how much of a language sits in public repositories, which is the gap IBM filled by training its own models on RPG.
What GitHub Copilot does
It helps you write code. GitHub describes it as an assistant that suggests code as you type, answers questions about a codebase, reviews your changes and works on tasks you hand it. You pick which AI model answers you, from a list that includes models by OpenAI, Anthropic, Google, Microsoft, Moonshot AI and xAI.
- Suggests the next lines of code while you type
- Answers questions about code you did not write
- Explains an existing program in plain language
- Reviews a change or a pull request
- Agent mode reads a repository, proposes a plan and edits files
Where it runs
- GitHub.com
- Visual Studio Code
- Visual Studio
- JetBrains IDEs, the integrated development environments from JetBrains
- Eclipse
- Xcode
- The Copilot command line tool
Visual Studio Code is the one that matters on IBM i. That is the same editor the free Code for IBM i extension uses to reach a system over SSH, the secure remote login service, so Copilot can sit in the window where your source members are already open. IBM went further with IBM Bob, which is built as its own fork of Visual Studio Code.
Does it know RPG?
GitHub does not say. Its documentation says inline suggestions support many programming languages, with quality that varies by the volume and diversity of training data for each language, and that a language with less representation in public repositories is harder to assist with. JavaScript is the example GitHub gives of a well-represented language. RPG, short for Report Program Generator, is not named either way.
IBM treated RPG as a special case rather than a language a general assistant would pick up on its own. IT Jungle reported on June 2, 2025 that IBM had fed more than 10 million lines of RPG into its Granite models and wanted more old RPG II and RPG III code. IBM i chief technology officer Steve Will said RPG is different enough that there were things IBM had to do differently and train differently.
How does it compare with IBM Bob?
GitHub Copilot and IBM Bob side by side
| GitHub Copilot | IBM Bob | |
|---|---|---|
| Made by | GitHub | IBM |
| Built as | An add-on for several editors | Its own fork of Visual Studio Code |
| Tuned for | Public code, many languages | RPG, CL, DDS and SQL |
| IBM i tools | None published by GitHub | About 30, in the Bob Premium Package for i |
| Price a month | $0 to $100 | $20 to $40 add-on, up to $240 with Ultra |
IBM Bob figures come from this glossary's IBM Bob and Bob Premium Package for i entries. Bob prices are reported figures, not an IBM list price.
GitHub Copilot plans, US dollars a month
Prices read from GitHub plan documentation on October 19, 2026. Copilot Free caps code completion at 2,000 suggestions a month. Verified students get the full features at no cost.
How Copilot and IBM i coding help lined up
- Jun 29, 2021GitHub launches Copilot as a technical preview, powered by OpenAI Codex. It works best on Python, JavaScript, TypeScript, Ruby and Go.
- Jun 21, 2022Copilot becomes generally available to individual developers at $10 a month or $100 a year.
- Jun 2, 2025IT Jungle reports IBM has fed more than 10 million lines of RPG into its Granite models and needs more.
- Mar 24, 2026IBM Bob reaches general availability, giving IBM i a coding agent tuned for its own languages.
- Oct 19, 2026GitHub lists Copilot Free plus five paid plans, $10 to $100 a month, and models from six vendors.
Copilot dates come from GitHub announcements. The RPG training figure and the Steve Will quote come from IT Jungle. The IBM Bob date comes from this glossary's IBM Bob entry.
Related Terms
IBM Bob (Project Bob)
IBM's generally available AI coding agent for IBM i, built as a fork of Visual Studio Code and tuned for RPG, CL, DDS, and SQL. Reached GA on March 24, 2026; Bob 2.0 and the IBM i-specific Bob Premium Package for i shipped June 24, 2026.
Bob Premium Package for i
The IBM i-specific add-on to IBM Bob, generally available June 24, 2026. It gives Bob about 40 skills and workflows and about 30 tools that work directly against QSYS, the IFS, CL, SQL and Db2 for i on a live system.
Code for IBM i
Code for IBM i is the free, open source Visual Studio Code extension that lets a developer browse libraries and edit and compile RPG, CL, COBOL and C source on an IBM i over SSH. It owns the connection to the system, so other extensions plug into it rather than building their own, and IBM Bob Premium Package for i installs it automatically, which makes it the layer AI coding help on IBM i actually runs on.
watsonx Code Assistant for i
IBM's November 2025 preview announcement of an AI code assistant purpose-built for RPG on IBM i. The preview track was superseded by IBM Bob, which reached general availability in March 2026.
RPG
Report Program Generator, the primary business application programming language on IBM i and its AS/400 predecessors, still running the majority of production logic in most IBM i shops.
Free-Format RPG
A modern RPG coding style that drops the rigid column-based syntax of traditional fixed-format RPG in favor of freer, more readable statement syntax.
IBM Granite
IBM's family of open-weight foundation models, released under the Apache 2.0 license, that underpins watsonx.ai and serves as the base model fine-tuned for RPG and IBM i in watsonx Code Assistant for i and IBM Bob.
Large Language Model (LLM)
A machine learning model trained on large volumes of text, used as the underlying technology behind generative AI tools including IBM Bob and general-purpose assistants like ChatGPT and Claude.
Generative AI
AI systems that generate new content, such as code, text, or explanations, rather than only classifying or scoring existing data.