Grafana k6 2.0 Brings AI-Assisted Performance Testing to Dev Teams
Grafana has officially released k6 2.0, a significant milestone for one of the most widely used open-source performance testing tools. With over 30,000 stars on GitHub, k6 has long helped engineering teams catch performance issues early. Now, k6 2.0 AI-assisted performance testing takes that mission even further.
The release builds directly on k6 1.0, which introduced TypeScript support, native extensions, and production-grade stability. This time, Grafana focuses on automation, AI integration, and making tests easier to author and scale.
At the heart of k6 2.0 are four new commands designed for AI-driven workflows. The k6 x agent command helps developers bootstrap agentic testing setups inside AI coding tools like Claude Code, Codex, and Cursor. Meanwhile, k6 x mcp exposes k6 through a built-in Model Context Protocol server, so AI agents can validate and run scripts directly. Developers also get k6 x docs for CLI-based access to k6 documentation and k6 x explore to browse the extension registry without leaving the terminal.
Together, these commands represent a shift in how k6 fits into modern development. Rather than just running scripts, k6 2.0 AI-assisted performance testing embeds testing deeper into the agent-driven workflow.
The extensions ecosystem also sees meaningful upgrades. Grafana now offers a consolidated catalog of both official and community extensions, making it easier to discover support for databases, message queues, DNS, streaming APIs, and more. New official extensions include k6/x/faker, k6/x/mqtt, k6/x/sql, and k6/x/dns. Community extensions like k6/x/sse and k6/x/kafka are also included. In addition, the xk6 tool evolves from a build utility into a full extension development toolbox, with commands like xk6 new, xk6 lint, and xk6 test to scaffold, validate, and publish extensions.
On the browser side, k6 2.0 greatly expands compatibility with Playwright, making it easier for teams to migrate existing Playwright tests to k6. Teams can now move naturally from functional correctness checks to full performance validation under load. The new Assertions API introduces an expect() syntax inspired by Playwright, supporting both non-retrying assertions for static values like HTTP status codes and auto-retrying assertions for dynamic browser elements.
Finally, k6 2.0 adds a JSON summary output so CI/CD systems and AI agents can read structured test results. Native OpenTelemetry output makes it simpler to align k6 metrics with existing application telemetry. For teams requiring production-scale load, k6 Operator 1.0 is now stable, enabling distributed testing on Kubernetes environments.
