📄 中文摘要
苹果的神经引擎在 A17 Pro 上每秒可处理 35 万亿次操作。大部分计算能力在用户支付月费使用他人服务器时未被充分利用。Off Grid 是一款免费的开源应用，能够直接在 iPhone 上运行大型语言模型。用户在首次下载后无需互联网连接，也不需要 iCloud 或苹果智能技术，仅需手机和模型即可使用。
如何在 2026 年的 iPhone 上本地运行大型语言模型（完全离线，无需订阅）
📄 English Summary
How to Run LLMs Locally on Your iPhone in 2026 (Completely Offline, No Subscription)
Apple's Neural Engine can process 35 trillion operations per second on the A17 Pro, with much of that power going unused while users pay monthly subscriptions to access someone else's server. Off Grid is a free, open-source app that allows large language models to run directly on your iPhone. After the initial download, no internet connection is required, and there is no need for iCloud or Apple Intelligence. Users only need their phone and a model to operate the app.