iPhone 17 Pro Demonstrated Running a 400B LLM
twitter.com
March 23, 2026
1 min read
🔥🔥🔥🔥🔥
71/100
Summary
Anemll successfully runs a 400 billion parameter model on an iPhone, achieving a performance of 0.6 tokens per second. The post has garnered over 113,000 views and numerous replies.
Key Takeaways
- Anemll successfully ran a 400 billion parameter model on an iPhone.
- The model achieved a processing speed of 0.6 tokens per second.
- The tweet received over 113,000 views and numerous replies.
Coverage & discussions
Hacker News · twitter.com
Community discussions
Community Sentiment
Positives
- Running a 400B parameter model on a handheld device showcases the incredible advancements in AI capabilities, making powerful tools more accessible.
- The potential for future models to achieve 100 t/s indicates rapid progress in both AI model efficiency and hardware development, which could revolutionize mobile AI applications.
Concerns
- The reliance on a mixture of experts raises questions about the actual performance and efficiency of the model, as not all parameters are active simultaneously.
Related Articles
GigaToken: ~1000x faster Language model tokenization
Jul 22, 2026
Harness Engineering 101: The LLM is the brain, the harness is the body
Oct 19, 2026
If you don't run your own LLM proxy, how do you know whats going on?
Oct 18, 2026
Hybrid Reasoning Agents: LLM agents that think in latent space, act through text
Oct 18, 2026
Llama.cpp on a jailbroken PS5: run any GGUF model natively on the console
Oct 16, 2026