Copilot stopped being an autocomplete two years ago, but 'preview' let enterprises dodge hard questions. GA is GitHub saying the agents are production-grade — now the security teams get to find out.
GitHub announced general availability of Copilot's autonomous coding agents, which can plan, write, test, and fix code across multiple files and open pull requests without human hand-holding. The feature ships to all Copilot plans today, from the $10 Pro tier to Enterprise.
GitHub today moved Copilot's autonomous coding agents from preview to general availability, bringing the feature to every paid Copilot plan across VS Code, Visual Studio, JetBrains IDEs, GitHub CLI, and Copilot Workspace. Agents can plan a change, edit across multiple files, run builds and tests, fix what breaks, and open a pull request with issue references and a summary of the work — all from a single instruction. The GA release adds concurrency tiers (five parallel agents on Pro, ten on Business, twenty on Enterprise), org-level policies, required code-review settings, and audit logs that are enabled by default for enterprise workspaces. GitHub says the preview already saw agents open more than fifteen million pull requests. GitHub is also retiring the older preview toggle and folding its documentation into the main Copilot docs, with a migration guide for teams that customized preview behavior.
Copilot began as a glorified autocomplete, and its evolution into an agent that finishes the job changes what 'developer platform' means. GA is more than a status badge: it forces GitHub to stand behind availability, pricing, and support in a way preview features never do, and it signals that agentic coding is now the default expectation in the tools developers already use. The stakes are existential for the company's roadmap — if agents genuinely do the grunt work, GitHub's pull request and review model becomes the control layer for AI output, not just human collaboration. That is the position GitHub has been quietly building toward for two years. The quiet pivot is that GitHub is now selling the review process as the product, framing human judgment as the safety layer around AI throughput.
GitHub says more than twenty million developers now use Copilot across its plans, and that surveyed enterprise teams using agents reported a median 40% reduction in pull-request turnaround time during the preview. Pricing is unchanged: Pro at $10, Pro+ at $39, Business at $19, and Enterprise at $39 per user per month, with agent runs counted against each plan's concurrency rather than billed per task. GitHub claims agents complete roughly two-thirds of coding subtasks within three passes, and that the required code-review setting is on by default for Business and Enterprise. The company also published an enterprise adoption guide, a telling addition for a feature that used to be sold on vibes. GitHub's own telemetry shows agents refactoring and deleting code more aggressively than humans typically do, which is both the power and the governance risk.
General availability lands in the middle of a crowded agentic-coding market where Cursor, OpenAI's Codex, and Devin are all fighting for the same developer mindshare, and GitHub's advantage is distribution: agents ship inside the repository where work already happens. For engineering leaders, the release turns an evaluation decision into a procurement decision, complete with SLAs, security baselines, and a per-seat price that suddenly looks cheap compared with dedicated agent tools. The competitive response will likely be a pricing war in agentic features across the category, with GitHub's bundled approach setting the floor. The real question is whether orgs measure agent productivity in pull requests merged, bugs escaped, or hours returned. Analysts estimate agentic coding tools now touch a meaningful slice of new production code, making this less a feature launch and more a platform statement.
Watch how regulated industries respond, because audit logs and required review are only convincing if they survive real compliance scrutiny, and that is where agent adoption will actually stall or spread. Also track whether GitHub's concurrency limits push heavy users into Enterprise, a classic monetization tell, and whether independent benchmarks show agent quality actually holding up in production repos. The developer-relations question matters too: communities that already distrust AI-written code will put the merge-rejection rate under a microscope. Finally, watch the next feature drop, because GA usually arrives right before the roadmap's bigger reveal.
This story was reported from primary materials: the companies' own documentation, on-the-record statements and data we could independently check. Numbers were re-verified against original sources rather than secondary aggregation, and analyst commentary is labeled as commentary — not reporting. Where we could not confirm a detail, we said so in the text instead of hedging vaguely.
Have context or a correction? Our news desk updates stories in place, with the change noted at the top of the article. Follow the StackHK news feed for the follow-ups as this story develops.
Three signals are worth watching from here: whether early adopter sentiment holds past the honeymoon window, whether pricing or packaging shifts to convert attention into revenue, and how direct competitors respond — in this category, answers usually arrive within weeks rather than quarters. As always with fast-moving AI news, the second-day story is often more consequential than the launch-day headline, and we will keep this article updated as the picture firms up.
Copilot's coding agents were never the surprise — the surprise is that GitHub is now comfortable calling them done.
“GA effectively ends the demo-ware debate; the real test is whether agent PRs survive code review at scale in regulated organizations.”
“The unremarkable milestone that changes everything — teams now assume agents will do the grunt work.”