iOS 27: Apple will let users pick AI models instead of its own

Apple is changing its strategy: in iOS 27, users will be able to choose which AI models to use instead of the built-in Apple Intelligence.

Freedom of choice

Rather than relying only on built-in AI features, users will be able to integrate third-party models. That means they can connect whichever solution suits a given task best: one tool for writing, another for analysis, a third for creative work. This approach gives users full control over which AI services they entrust with their data. It would also let them adopt the newest and most powerful models as they are released, without waiting for an iOS update.

Which models can be connected

The main players are expected to join the ecosystem:

• OpenAI (ChatGPT, GPT-5.5)
• Google (Gemini)
• Anthropic (Claude)
• Meta (Llama)*
• Other developers of independent models

Users will be able to pick models for different types of tasks, for example one for coding, another for text analysis, a third for creative writing. This turns the iPhone into a multi-tool device rather than a one-size-fits-all system.

Why Apple is doing this

The decision may seem surprising, but it is logical. Apple cannot simultaneously outdo OpenAI, Google and Anthropic, all of which have invested billions in developing their models. Instead of competing on quality, Apple is positioning itself as the platform on which that competition takes place.

It is also a response to antitrust pressure. By giving users a choice, Apple shows that it is not trying to lock out competitors. The strategy works in Apple's favor: it remains an intermediary and a platform, even if its own AI development lags behind.

What this means

For the market, it is a signal that the era of built-in, closed AI assistants is ending. Companies with the best model and the best integration win. For users, it is a chance to build a truly personalized AI stack tailored to their needs.

*Meta has been designated an extremist organization and is banned in Russia.