The cleanest way to find out whether an iPhone AI app is actually "offline" is to turn on Airplane Mode and ask it something. The ones that answer are on-device. The ones that fail with a network error are cloud apps that mislabeled themselves. The ones that silently replace your prompt with "I need an internet connection to help with that" are the most annoying: they pretend to be apps and they're really just wrappers.
We ran the airplane-mode test on ten popular iPhone AI apps. Five pass. Three fail cleanly. Two fail in ways that would annoy you on a flight. This is the ranking. PocketLLM is #1 because it was designed from day one to pass this test.
The test
Every app below was installed on the same iPhone 15 Pro, given a reasonable chance to download a model if that was part of its setup, and then tested with Airplane Mode on (Wi-Fi off, Cellular off, Bluetooth on so we could pair a keyboard). The test prompt, for every app: "Write a three-sentence summary of the French Revolution." We noted whether the app answered at all, how fast, and how coherent the output was. Then we asked a second, harder prompt: "Explain quantum entanglement to a 10-year-old." Same criteria.
The 10 apps, ranked
1. PocketLLM — Passes cleanly
Designed from day one for airplane mode. Runs Llama 3.2 3B and other models directly on-device. Responded instantly, with quality comparable to ChatGPT Free on both test prompts. No network indicator. Nothing in the UI even hints at connectivity. This is how an offline AI app should feel. Currently in waitlist / early access.
2. Private LLM — Passes cleanly
Native iOS app, fully on-device, paid. Same test passed with the same quality as PocketLLM. Notably, it works exactly the same in airplane mode as in normal mode — which is the point. The app makes no network calls during inference, period.
3. LLM Farm — Passes cleanly
Open-source, free, fully on-device. Requires a bit more setup (you pick the model, you manage quantization) but once configured, airplane mode is entirely transparent. Passed the test. Slightly slower generation than PocketLLM or Private LLM but equally reliable.
4. MLC Chat — Passes cleanly
Research-grade, from the MLC-LLM project. Free, open source. Passed the test. UI is less polished than the top three but the underlying runtime is rock solid and the airplane-mode behavior is identical to the others.
5. Apple Intelligence (on iPhone 15 Pro+) — Passes the easy prompts, fails the hard ones
Apple Intelligence runs on-device for simpler requests and tries to fall back to Private Cloud Compute for harder ones. In airplane mode, the on-device path works — summarization, writing tools, simple Q&A. The harder prompts that would normally hit Private Cloud Compute either fail silently or return a shorter on-device response. For "some things work, some don't" airplane use, it's fine; for "assume offline for everything," the other four are more predictable.
6. ChatGPT — Fails cleanly
Network error immediately. No cached responses. No offline mode. The app itself launches and the UI loads, but any prompt results in a connection error message. To ChatGPT's credit, the error is clear — you know immediately that you're out of luck.
7. Claude — Fails cleanly
Same as ChatGPT. Network error, no cached offline capability, no pretense. Anthropic's app just tells you it can't connect. This is the correct behavior for a cloud-only product, even if it's disappointing in an airplane.
8. Google Gemini — Fails cleanly
Network error. No on-device fallback (despite Google shipping Gemini Nano on some Android devices; on iOS the Gemini app is fully cloud). Fails the same way ChatGPT and Claude do.
9. "Private AI Chat" (App Store clone) — Fails misleadingly
We tested one of the top App Store results for "private AI chat" that described itself as "offline" in its marketing. In airplane mode, the app loaded, accepted a prompt, and returned a generic apology about "no internet." It is a cloud app that was marketed as offline. We're not naming it because the names change weekly — the pattern is the issue. If you find an iPhone AI app in the App Store marketing itself as "private" or "offline," test it in airplane mode before trusting it.
10. Character.AI — Fails loud
Character.AI on iPhone is a thin client for a cloud service. In airplane mode it fails immediately with "no connection," which is expected — Character.AI has never claimed to work offline. Included in the test as a baseline cloud app to confirm the test methodology.
The results table
| # | App | Airplane mode? | Why |
|---|---|---|---|
| 1 | PocketLLM | Passes | Fully on-device by design |
| 2 | Private LLM | Passes | Fully on-device |
| 3 | LLM Farm | Passes | Fully on-device, open source |
| 4 | MLC Chat | Passes | Fully on-device, research-grade |
| 5 | Apple Intelligence | Partial | On-device path works, PCC doesn't |
| 6 | ChatGPT | Fails | Cloud-only, clean error |
| 7 | Claude | Fails | Cloud-only, clean error |
| 8 | Google Gemini | Fails | Cloud-only on iOS |
| 9 | "Private AI Chat" clone | Fails (misleading) | Cloud under "offline" marketing |
| 10 | Character.AI | Fails | Cloud-only by design |
PocketLLM is currently in waitlist / early access.
How to test any app yourself
Before you rely on an iPhone AI app for a flight, test it. Swipe down from the top-right corner of your iPhone, tap Airplane Mode, and ask the app a simple question. If it answers, it's on-device (or has cached responses — which is uncommon but worth knowing). If it errors cleanly, it's cloud-only and will fail on the flight. If it returns something weird — a generic apology, a "let me try again later" message, a fallback response — it's doing something hacky and you should not trust its privacy claims either.
The airplane-mode test is also the single best privacy test. If an app can't answer without the network, then by definition your prompts are transiting someone's server. "Private" and "works in airplane mode" are effectively the same category.
The quick answer
The iPhone AI apps that actually work in airplane mode in 2026 are PocketLLM, Private LLM, LLM Farm, and MLC Chat — the four native on-device apps. Apple Intelligence works partially. Everything else is cloud-only and fails. Test any "offline" app yourself with airplane mode before trusting the marketing.
For the full on-device iPhone ranking beyond just airplane mode, see our best on-device LLM apps for iPhone roundup.