Apple to Let Users Pick Third-Party AI Models

Apple is preparing a landmark shift in its artificial intelligence strategy by allowing users to select from a range of third-party AI services to power key software features across its devices. The change, expected to arrive this fall with iOS 27, iPadOS 27, and macOS 27, signals Apple's ambition to transform its hardware ecosystem into a comprehensive AI platform rather than relying solely on its own models.

According to people familiar with the matter, users will be able to choose among multiple external AI models for tasks such as generating and editing text and images — a move that could reshape how over 1 billion Apple device owners interact with artificial intelligence on a daily basis.

Key Takeaways at a Glance

• Apple will offer users a choice of multiple third-party AI models within its operating systems
• The feature targets core tasks including text generation, text editing, and image creation
• Expected rollout timeline: fall 2025 alongside iOS 27, iPadOS 27, and macOS 27
• The strategy positions Apple as an AI platform aggregator rather than a single-model provider
• This builds on Apple's existing partnership with OpenAI's ChatGPT, integrated into Siri since late 2024
• Competing AI providers like Google Gemini and Anthropic Claude could gain direct access to Apple's massive user base

Apple Embraces a Platform-First AI Strategy

This announcement represents a significant philosophical pivot for a company historically known for its walled-garden approach. Unlike Google, which pushes its own Gemini models across Android and Search, or Microsoft, which has bet heavily on OpenAI's technology, Apple appears to be positioning itself as a neutral marketplace for AI capabilities.

The strategy mirrors what Apple has done successfully in other domains. The App Store, for instance, turned the iPhone into a platform where third-party developers — not Apple — provided the majority of user value. Now Apple seems to be applying the same playbook to artificial intelligence.

By offering choice, Apple also sidesteps a growing concern among consumers and regulators: AI vendor lock-in. Users who prefer one model's writing style or another's image generation capabilities will have the freedom to switch without leaving Apple's ecosystem. This is a competitive advantage that neither Google nor Samsung currently offers on their respective platforms.

How This Builds on Apple Intelligence

Apple first unveiled its AI framework, branded as Apple Intelligence, at WWDC 2024. The initial rollout included on-device models for summarization, notification management, and writing tools, alongside a partnership with OpenAI to integrate ChatGPT directly into Siri for more complex queries.

However, that first iteration drew mixed reviews. Critics noted that Apple Intelligence felt limited compared to standalone AI tools, and the ChatGPT integration — while useful — created an awkward experience where users were prompted to leave Apple's native interface for certain tasks.

The new multi-model approach addresses these shortcomings directly:

• Deeper integration: Third-party models will power features natively within Apple apps, eliminating the need for separate prompts or redirects
• Broader capability: Different models excel at different tasks — one might be superior at creative writing while another leads in code generation
• Competitive pricing: By creating a marketplace dynamic, Apple could drive AI providers to offer better performance at lower costs
• User autonomy: Consumers gain control over which AI 'brain' powers their device, a unique selling proposition in the current market

Which AI Providers Could Join the Platform?

While Apple has not officially confirmed specific partners, industry analysts point to several likely candidates based on existing relationships and market positioning.

OpenAI is the most obvious inclusion, given its existing ChatGPT integration with Siri. The company's GPT-4o and upcoming models would likely be available as a default or premium option. Google, despite being a direct competitor in hardware, has shown willingness to license its Gemini models broadly — and accessing Apple's 1.5 billion active devices would be an enormous distribution win.

Anthropic, the maker of Claude, has positioned itself as the safety-focused alternative and could appeal to enterprise users and privacy-conscious consumers. Meta's Llama models, which are open-source, present an interesting possibility for on-device processing that aligns with Apple's privacy-first messaging.

Smaller players like Mistral AI, the French startup valued at over $6 billion, and Perplexity, known for its search-oriented AI, could also see this as a breakthrough distribution channel. For these companies, a spot on Apple's platform could be worth more than years of independent marketing spend.

The Business Model Behind the Shift

Apple's motivations extend well beyond user experience. The company has been facing pressure from investors to demonstrate a compelling AI strategy that can drive its next growth cycle, particularly as iPhone upgrade cycles have slowed.

A multi-model AI marketplace opens several revenue streams:

• Revenue sharing: Apple could take a percentage of subscription fees for premium AI models, similar to its 15-30% App Store commission
• Hardware differentiation: More powerful AI features create stronger reasons to upgrade to newer iPhones, iPads, and Macs with advanced Neural Engine chips
• Services growth: AI model subscriptions could become a new pillar alongside Apple Music, iCloud+, and Apple TV+, boosting the company's high-margin services segment
• Enterprise expansion: Offering model choice makes Apple devices more attractive to businesses that have standardized on specific AI providers
• Data leverage: While Apple processes AI on-device where possible, routing cloud-based AI requests gives the company valuable insight into usage patterns

Apple's Services division generated over $96 billion in revenue in fiscal 2024. Adding AI model subscriptions to that portfolio could meaningfully accelerate growth, especially if Apple negotiates favorable terms with providers eager for distribution.

What This Means for Developers and Businesses

For app developers, the implications are substantial. If Apple provides standardized APIs that allow apps to tap into whichever AI model a user has selected, developers could build more intelligent features without committing to a single AI provider. This reduces dependency risk and development costs.

Businesses already deploying Apple devices across their workforce — and there are millions of them — gain flexibility to align their AI tools with corporate policies. A financial services firm might mandate Claude for its safety guardrails, while a creative agency might default to a model optimized for visual content.

The move also raises questions about Apple's own AI development. The company has invested billions in its on-device models and custom silicon. Rather than abandoning that work, Apple appears to be layering third-party options on top of its foundation models, creating a tiered system where basic AI tasks run locally and more demanding requests are routed to the user's chosen cloud provider.

Industry Context: The Platform Wars Heat Up

Apple's decision arrives at a pivotal moment in the AI industry. Microsoft has deeply embedded OpenAI's technology into Windows through Copilot, making it the default AI experience for PC users. Google has woven Gemini into Android, Chrome, and Search. Samsung has partnered with Google for its Galaxy AI features.

Each of these competitors has chosen a single-provider strategy. Apple's multi-model approach is a deliberate counter-positioning that could prove decisive. In a market where no single AI model dominates across all tasks, giving users choice may be the most pragmatic — and most 'Apple' — solution.

The timing also coincides with increasing regulatory scrutiny of AI monopolies in both the EU and the United States. By offering model choice, Apple preemptively addresses concerns about anti-competitive bundling that have plagued its App Store practices.

Looking Ahead: Fall 2025 and Beyond

The official unveiling is widely expected at WWDC 2025 in June, with the public launch following in September or October alongside new iPhone hardware. Key questions remain about implementation details — including whether all AI providers will be free, whether Apple will curate the available models, and how deeply third-party AI will integrate into system-level features like Siri.

If executed well, this strategy could establish Apple as the definitive AI platform for consumers, much as it became the definitive app platform with the iPhone. The company would not need to win the AI model race — it would simply need to be the best place to use whichever model wins.

For the broader industry, Apple's move validates a critical insight: in the age of AI, distribution matters as much as model quality. The companies that control where and how users access AI will wield enormous influence over which models succeed — and Apple, with its unmatched consumer hardware ecosystem, may hold the strongest hand of all.