#siri

Apple has seeded the release candidate of iOS 27.2 to registered developers, a week after the fourth beta of the update. An iPadOS 27.2 release candidate arrived alongside it, according to MacRumors.

A release candidate is normally the build Apple intends to ship, provided testers do not turn up a serious bug. MacRumors says it expects iOS 27.2 to reach the public before the end of October, possibly as soon as next week. Apple has not given a release date.

The headline addition is expanded Siri AI language support. Apple had previously said the new languages would arrive in October, and MacRumors lists French, Japanese, Korean, Portuguese and Spanish among them.

Apple has confirmed that Siri AI will have a daily usage limit, and that limits will also apply to other Apple Intelligence features. The details appear in Apple's published terms, which MacRumors read.

When a person reaches a limit, the feature becomes available again after a time-out. Apple says the limits can vary by feature, by the complexity of a request, by system demand and by system policies, so there is no single number users can expect.

Apple also says expanded access will be offered for a fee at some point. It has not said how much this will cost or how it will be billed. At WWDC in June, Apple indicated that higher iCloud+ tiers would include expanded access to the limited Apple Intelligence features.

Apple said at its September event that iOS 27 will be released on Monday, September 14, together with iPadOS 27, watchOS 27 and macOS 27. It is the public debut of the rebuilt Siri, which Engadget says Apple developed in partnership with Google using Gemini models.

Engadget's early testing found that the new Siri could answer complex questions and handle multi-step requests. On the Mac, Siri AI is built into Spotlight.

Compatibility is a big part of the story. Siri AI supports the iPhone 15 Pro and newer, so owners of older iPhones will not get it. The iPhone 17, 17 Pro Max and iPhone Air get a better experience thanks to a more capable on-device model that improves speech recognition and dictation.