AIHacker Newsabout 2 hours ago
Qwen3.8 Runs on iPhone 18 Neural Engine
1 min read
User Anemll leaked details showing Qwen3.8-27B model running on the iPhone 18 Neural Engine with 24K context. The system achieves 70 tokens per second at 1.5 bits per word. A draft model reaches 2.8 bits per word using mixed vector quantization. This performance data appears on a forum post dated October 17, 2026.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyApple
