Apple is reportedly closing in on a deal with Google that would bring a custom version of the Gemini AI model to Siri, with Apple paying around $1 billion per year. Citing a new report from Bloomberg, the agreement would help power a major Siri overhaul and a wave of AI-driven features while Apple continues developing its own in-house models.
This partnership marks a strategic shift for Apple, which typically leans on its own technology. The plan, as reported, is to use Google’s model as a temporary solution until Apple’s homegrown AI becomes strong enough to take over. The goal is to give Siri a significant boost in understanding, responsiveness, and capability across iPhone and other Apple devices.
One of the most striking details is the size of the custom Gemini model involved: about 1.2 trillion parameters. Parameters are a key measure of an AI model’s complexity and potential capability. By comparison, the current cloud-based version of Apple Intelligence uses around 150 billion parameters, making the proposed Gemini setup roughly eight times more complex. In practical terms, that kind of scale could enable more natural conversations, better context handling, and more reliable assistance for complex tasks.
Before leaning toward Google, Apple reportedly evaluated models from OpenAI and Anthropic earlier this year. After testing all three, Apple chose to move forward with Google’s offering for the Siri revamp.
The upgraded Siri experience is expected to roll out as soon as next spring, though the timeline and feature set could still shift as work continues. If finalized, the deal would signal a major step in Apple’s AI strategy, combining a near-term performance leap for Siri with the longer-term ambition of transitioning to Apple’s own advanced models.
Key takeaways for users: Apple is aiming to make Siri faster, smarter, and more helpful; the company is leveraging a custom Gemini model to accelerate improvements; and while the initial rollout is targeted for next spring, plans may evolve before launch.
