Apple’s next-generation Siri will debut next month! iOS 26.4 beta version released first.

Fast Technology reported that earlier this month, Apple and Google officially announced a partnership to provide artificial intelligence support for product features (such as Siri) to be launched later this year.

According to a recent report by renowned journalist Gurman, the results of this collaboration will be revealed as early as next month in the iOS 26.4 beta version, and Apple also plans to demonstrate the relevant features to the public in some form.

It is understood that the next generation of Siri will be powered by Google’s Gemini model and cloud technology.

Apple plans to pay Google approximately $1 billion annually for the right to use a customized Gemini model, with the core version being the Gemini 2.5 Pro with 1.2 trillion parameters.

This model will be responsible for Siri’s core functions such as message summarization, task planning, and execution of complex multi-step commands. Compared to the current cloud model with 150 billion parameters used by Siri, its processing power and depth of contextual understanding have achieved a leap in magnitude, and it supports multimodal interaction and ultra-long text processing of 128K tokens.

The upgraded Siri is expected to be released with iOS 26.4 in the spring of 2026, and will be further integrated into iOS 27, macOS 27 and other systems previewed at WWDC in June 2026.

It is understood that Apple previously tested three mainstream models: Google Gemini, OpenAI ChatGPT, and Anthropic Claude. The core reasons for ultimately choosing Google include: better financial costs, a long-term cooperative relationship between the two parties, and Gemini 2.5 Pro’s current leading performance in the large language model leaderboard.

Gemini employs a hybrid expert architecture. Although the total number of parameters exceeds one trillion, only a portion of the parameters are activated for each query. This ensures high performance while reducing computing costs and energy consumption, making it compatible with multiple device scenarios such as iPhone and Mac.

The AI ​​model that Google developed for Apple will run on Apple’s private cloud computing servers, preventing Google from accessing Apple user data and alleviating users’ concerns about data breaches.