GitHub: what changes for developers building with agents

At GitHub Universe on 28 October 2026, GitHub's Jay Parikh framed the shift plainly: developers have moved from writing every line of code to orchestrating fleets of agents. The stated goal is completed software work, not more code.

GitHub presented a Copilot app built for that workflow. It connects agents and tools across projects, and it also handles specs, design discussions and telemetry. Copilot projects run agent work across several repositories in parallel managed sessions, available natively in the app from that day.

Session continuity lets one cloud agent session follow you across the GitHub app, VS Code, the CLI and the web. The agent keeps running in the cloud while you switch surfaces.

Different tasks suit different models, so GitHub introduced Copilot Fusion. It assesses what a task needs across reasoning, code generation, debugging and tool use, then picks how to run it across one or more models.

GitHub reports that Fusion Max scored within two points of Opus 5.5 Max on the Artificial Analysis Coding Agent Index. It says that came at half the cost and 40% less time per task. Those are GitHub's own figures, not an independent test.

On review, Copilot code review Max posted the highest F1 score on ReviewBench for high-severity issues, according to GitHub. Lite and Balanced remain the faster everyday options.

From 28 October 2026, GitHub Enterprise customers can use all Copilot features without buying a separate subscription. That covers the Copilot app, code review and automations. You pay for the AI credits you consume instead.

Availability matters here: GitHub says this runs through GitHub Sales, with self-serve coming soon. Existing Copilot Business and Copilot Enterprise customers can keep their current plans or switch when ready.

Automations can now learn from their outcomes and propose improvements for you to review and approve. GitHub still lists Cloudflare, Datadog and Figma among the tools that keep working alongside it.

GitHub's Octoverse report, also published on 28 October 2026, counts 17.9 billion pushes in the previous 12 months, more than 2024 and 2025 combined. Commits reached 34.3 billion and merged pull requests 1.3 billion.

Among repositories created this year, the average one with an agent configuration file logged eight times as many commits as one without. As of September, developers started more than 40% of pull requests with an agent, up from 17% in March.

GitHub also reports 225 million developers on the platform. The United Kingdom appears at 6.47 million in its September 2026 country table.

GitHub engineers Eleanor White and Victoria Dye described the infrastructure work in a post on 21 October 2026. Nearly all repositories now live in two places: GitHub's own data centres and Azure. The busiest ones were held back by a roughly 25ms round trip between the two.

The fix changes what the write quorum locks. Instead of serialising the whole repository, GitHub now serialises individual references. Writes to different branches no longer wait on each other.

In internal testing with realistic traffic, writes to different references saw 3–4x higher throughput at lower latency. That is GitHub's own measurement, not a public benchmark.

GitHub says it expects to complete its migration to a multi-region Azure deployment early next year. The per-reference write design is described as the bridge that gets the busiest repositories into Azure, and GitHub says it raises a repository's write throughput ceiling roughly in proportion to how many references are written at once.

The Copilot figures come from GitHub's own benchmarks and telemetry. Nothing here shows how the tools behave on your codebase or what your credit bill will look like. The self-serve route for enterprise Copilot is still listed as coming soon.

If you maintain a busy repository, the practical question is whether your team pushes to shared branches or separate ones. GitHub's engineers note that most writes touch different references, such as agents and developers pushing to their own branches, which is the case the new write model is meant to help most.