"Apple is preparing to power its AI-upgraded Siri with a custom version of Google's Gemini AI model, according to a report from Bloomberg's Mark Gurman. The iPhone maker will reportedly pay Google around $1 billion annually to gain access to the technology, which will generate summaries and perform planning-related tasks. As noted by Bloomberg, Apple will run the custom Gemini model on its Private Cloud Compute servers, though it will continue to use in-house models to power some Siri features."
"Gurman reports that the Gemini model will use 1.2 trillion parameters to help inform its responses, far greater than the 150 billion-parameter model behind the cloud-based version of Apple Intelligence. Apple has been scrambling to get back on track after the company delayed the rollout of new AI-powered features inside Siri. In June, Bloomberg reported that Apple had weighed using models from OpenAI or Anthropic to power the new Siri before reportedly settling on a partnership with Google."
Apple plans to integrate a custom Google Gemini model into an upgraded Siri and will pay roughly $1 billion per year for access. The Gemini model will run on Private Cloud Compute servers while Apple retains in-house models for some Siri capabilities. The custom Gemini is reported to use 1.2 trillion parameters, enabling generation of summaries and planning-related tasks and exceeding the 150 billion-parameter model used by Apple Intelligence. Apple delayed the Siri rollout, evaluated OpenAI and Anthropic options before partnering with Google, and continues internal AI development that could eventually replace the Gemini integration.
Read at The Verge
Unable to calculate read time
Collection
[
|
...
]