Apple has released visionOS 27.2 for Apple Vision Pro, and it is a bug fix and security update with no new features. You can install it from Settings, then General, then Software Update on the headset.
Installing the update requires taking the headset off, and a progress bar appears on the EyeSight display. It runs on every Apple Vision Pro that runs visionOS 27, which covers both the M2 and M5 models, and Apple has not yet published the security content for it.
visionOS 27.2 fixes bugs and adds nothing you will see
Apple calls it a bug fix and security update. The release candidate carried one known issue, where 3D objects opened from Safari may reposition unexpectedly when you move them.
The release candidate arrived on October 19 as build 24N118. visionOS has no public beta program, so only registered developers tested it first. Beta 1 arrived on September 16, two days after visionOS 27 launched. Beta 2 followed on September 21, beta 3 on October 5 and beta 4 on October 13. The last update before this one was visionOS 27.0.1 on September 28, build 24M372, which has no published security entries.
visionOS 27 shipped on September 14 with Siri AI in beta, app windows that curve around your field of view, Wi-Fi connections up to three times faster and panorama photos that turn into spatial environments. Vision Pro is one of two Apple platforms where Siri AI works in the EU at launch.
Some features, such as adjustable Siri voices and improved dictation, are limited to the M5 model.
Here are the full release notes from Apple:
This update provides bug fixes and security updates for your Apple Vision Pro. It also includes the following features and enhancements:
Siri AI
Excluded Apps lets you choose which apps Siri, Search, and suggestions can access, all from one setting
Granular text selection lets you select and copy just the part of a Siri response you need
Siri AI is available in more languages, with support for French, Japanese, Korean, and Spanish. Siri AI is not available in mainland China
Other
Support for SynthID watermarking to help identify whether an image was generated or modified with generative tools in Photos and Image Playground
FaceTime supports adding a third party sign language interpreter or live captioner to a call
