Release notes··3 updates
3 updates shipped on October 16, 2026.
Highlight
Featured this release
A new selectable web agent, Fast-v3, matches our top-tier pass rate on our benchmark set at a fraction of the cost.
Pick Fast-v3 ("Quick, cheap and efficient") as the agent for web tests. It's tuned to wait for slow-loading pages and passed 700 of 765 benchmark tests, the same as our Haiku-based agent.
More on this: web testing, AI testing and end-to-end testing.
Update 02 of 03
The coverage map shows your app's features as labelled hexagons linking the pages and actions they span.
Features appear as hexagon hubs on the coverage map, linked to the pages their actions live on. Each hexagon's outer ring fills with the share of tested actions, so well-covered features stand out. Hover a feature to highlight its actions and pages; click it to see "x tested of y" in the info panel. Hovering a page or action lights up every feature that touches it.
More on this: exploratory testing and CTOs.
Update 03 of 03
Reruns update the existing review, and run notifications now lead with the review verdict.
If a run already has a review chat and you rerun failed tests from the run page, the same conversation re-checks the latest results. The Analysis tab shows the previous review with a banner until the new one lands.
Slack, Teams and email notifications now lead with the verdict – for example "Needs attention: 2 failures to fix" – list the first five failures and setup issues, and link to "Open analysis" and "Open review chat". In the review, "Issues" is now called "Failures".
More on this: notifications testing and integrations.
QA.tech agents test your product autonomously, so moving fast never means shipping broken. See how it works in a 30‑minute demo.