Title: Running 400B Model on iPhone | BestBlogs.dev

URL Source: https://www.bestblogs.dev/status/2036296019835756929

Published Time: 2026-03-24 04:16:18

Markdown Content: Skip to main content Toggle navigation menu Toggle navigation menuArticlesPodcastsVideosTweetsSourcesNewsletters

⌘K

Change language Switch ThemeSign In

Narrow Mode

Running 400B Model on iPhone

Running 400B Model on iPhone

@simonw

看看这个：Qwen3.5-397B-A17B —— 一个 397B 的模型 —— 使用流式加载 MoE 权重的技巧，竟然能在 iPhone 上运行！

!Image 3: Anemll

#### Anemll

@anemll · 1d ago

Running 400B model on iPhone!

0.6 t/s

Credit @danveloper @Alexintosh @danpacary @anemllShow More

!Image 4: 视频缩略图

03:51

91

114

1,510

199.2K

Mar 24, 2026, 4:16 AM View on X

8 Replies

3 Retweets

50 Likes

6,602 Views  Simon Willison @simonw

One Sentence Summary

Demonstrates the scalability of the SSD-streaming MoE technique by successfully running a 397B parameter Qwen model on an iPhone.

Summary

This tweet showcases the impressive scalability of the SSD-streaming MoE technique. By applying the same method used on MacBooks, the author demonstrates that a massive 397B parameter Qwen model can be executed on an iPhone, further pushing the boundaries of what is possible on mobile hardware.

AI Score

84

Influence Score 10

Published At Today

Language

English

Tags

LocalLLM

iPhone

Qwen

Optimization

MobileAI HomeArticlesPodcastsVideosTweets

Running 400B Model on iPhone | BestBlogs.dev