GitHub Copilot
GitHub (Microsoft)GitHub Copilot's agent mode, which takes an assigned task and runs it as autonomous background work, planning, exploring the repository and executing changes. It orchestrates the first-party Copilot agent alongside Anthropic and OpenAI agents, and is built natively into GitHub across VS Code, Visual Studio, JetBrains and Neovim.
Scored as the agent surface; the in-IDE assistant is a separate record in ui_api. GitHub reports all-time users and no active-user figure. Verified 2026-08-13 via the GitHub Copilot product page.
Openness
1 high confidence- source
- closed
- license
- proprietary(GitHub/Microsoft SaaS)
- self-host
- no
Proprietary GitHub/Microsoft product; agent mode is a closed SaaS feature, model-agnostic on the backend but not open source.
- https://github.com/features/copilot recorded 2026-08-13
GitHub Copilot sold as a hosted proprietary product with agent mode and task assignment to agents, natively built into GitHub; no source distribution and no self-host option
Adoption
5 high confidenceGitHub Copilot has surpassed 20M+ all-time users, which Microsoft describes as "the world's most widely adopted AI developer tool"; agent mode is the autonomous surface of that base. That bands at >10M users, level 5. The 20M is an all-time cumulative count rather than an active one, and no agent-mode-specific figure is disclosed, so it stands in for an active count the product does not publish - which means the true active figure is lower, not higher.
- https://github.com/features/copilot recorded 2026-08-13
'growing to millions of individual users and tens of thousands of business customers, GitHub Copilot is the world's most widely adopted AI developer tool'
- https://techcrunch.com/2025/07/30/github-copilot-crosses-20-million-all-time-users/ recorded 2026-08-13
more than 20 million users from the Microsoft earnings call, confirmed by GitHub as "all-time users"; Microsoft and GitHub do not report monthly or daily actives
Capability
4 medium confidenceBroad agentic feature set with frontier-model backends and best-in-class repository integration: autonomous background agents that plan, explore and execute, multi-model orchestration across Copilot, Claude and Codex, and deep GitHub, pull-request and CI integration. No SWE-bench figure is published for Copilot agent mode itself, so this is a judgment on features rather than a measurement, strong but below where a verified benchmark could take it.
- https://github.com/features/copilot recorded 2026-08-13
agent mode in the editor, task assignment to agents, multi-model orchestration and IDE coverage across VS Code, Visual Studio, JetBrains and Neovim; no benchmark figure published
Verified 2026-08-13