TL;DR
Apple’s major Siri overhaul — first shown at WWDC 2024 — has been delayed while the company rebuilds its assistant on a new LLM-based stack. Reports now place the updated Siri and an Apple-built generative search feature on track for iOS 26.4, likely arriving in March or April with a beta expected next month.
What happened
At WWDC 2024 Apple outlined three headline upgrades for Siri under the Apple Intelligence umbrella: personal context (Siri can use data on your device to answer questions), on-screen awareness (Siri can interpret what’s on your screen during a request) and taking direct action inside apps. Those changes were originally slated for iOS 18.4 but were postponed after Apple encountered engineering problems and chose to rebuild Siri on a new large language model (LLM) infrastructure. According to reporting and uncovered code, Apple is now considering using Google’s Gemini model for the new assistant, with queries routed through Google’s Private Cloud Compute to limit data exposure. Separate reporting also describes an internal project called World Knowledge Answers (WKA), a generative AI search feature that may debut alongside the rebuilt Siri. Based on past release patterns, iOS 26.4 and its beta could arrive in the March–April window.
Why it matters
- New Siri features aim to reduce the need to locate content or open apps manually, simplifying common tasks.
- On-screen awareness and personal context could make voice and typed requests more conversational and contextual.
- Choosing a cloud LLM like Gemini would change the technical architecture and raise privacy and data-handling questions; Apple says Private Cloud Compute would be used.
- An Apple-built generative search product would position the company directly against third-party AI search tools such as Perplexity and ChatGPT.
Key facts
- WWDC 2024 introduced three planned Siri upgrades: personal context, on-screen awareness and taking actions in apps.
- Those Siri upgrades were initially expected in iOS 18.4 but were delayed for an engineering rewrite.
- Apple is rebuilding Siri on a new LLM-based infrastructure rather than the previous engineering stack.
- Apple is reportedly considering Google Gemini to power the new assistant, using Google’s Private Cloud Compute to protect query data.
- Macworld reporting and uncovered code indicate the new Siri features are being prepared for iOS 26.4.
- Mark Gurman’s reporting says Apple’s generative search project is internally called World Knowledge Answers (WKA).
- WKA is described as a generative AI search feature that could appear as soon as March, possibly alongside iOS 26.4.
- WKA may eventually expand into Safari and Spotlight, according to reporting, though timing and scope are not finalized.
- Based on historical release timing, iOS 26.4 is likely to appear in March or April, and its first beta could arrive next month.
What to watch next
- Timing and availability of iOS 26.4 betas and the public release (first beta could arrive next month).
- Whether Apple finalizes a decision to use Google Gemini for Siri’s LLM backing — not confirmed in the source.
- How and when World Knowledge Answers (WKA) might be integrated into Safari and Spotlight — not confirmed in the source.
Quick glossary
- Apple Intelligence: Apple’s umbrella term for built-in machine learning and AI features that augment iOS, macOS and other platforms.
- LLM (Large Language Model): A type of machine-learning model trained on large amounts of text to generate or interpret human language.
- Google Gemini: A suite of language models developed by Google that can be used to power generative AI features.
- Private Cloud Compute: A cloud-processing approach intended to run models without exposing raw user query data to the cloud provider.
- Generative AI search: A search experience that synthesizes answers using generative models rather than returning only ranked links.
Reader FAQ
When will the new Siri arrive?
Reporting places the rebuilt Siri and related features in iOS 26.4, which historically would land in March or April; a beta could appear next month.
Will Siri run on-device or in the cloud?
Apple is reportedly weighing use of Google Gemini rather than on-device models; if used, queries would run via Private Cloud Compute.
Will Apple’s generative search replace Safari’s current search?
The source says WKA may eventually be added into Safari and Spotlight, but specific plans and timing are not confirmed in the source.
Why was Siri delayed from the original schedule?
Apple encountered technical and engineering hurdles and chose to rewrite Siri’s underlying stack on a new LLM-based infrastructure.
Rumor Replay: iPhone Fold’s crease-free display, and more Ryan Christoffel Jan 8 2026 Here’s what to expect with Siri and Apple Intelligence in the coming months Michael Burkhardt | Jan 10…
Sources
- Here’s what to expect with Siri and Apple Intelligence in the coming months
- Everything Apple Is Releasing in 2026: iPhone Fold, LLM …
- What to expect from Apple in 2026, and when
- Apple's Much-Anticipated Siri Upgrade Is Almost Here
Related posts
- AI as a Business-Model Stress Test: Tailwind Layoffs and the Shift to Operations
- Yuanzai World: LLM-driven RPGs Let Players Shape Branching World-lines
- ChatGPT Health Is a Marketplace — Guess Who Becomes the Product?