iPhone 17 Pro Demonstrated Running a 400B LLM

Re:iPhone 17 Pro Demonstrated Running a 400B LLM by simopa on Monday March 23, 2026 @10:57AM (#47490431)

It's crazy to see a 400B model running on an iPhone. But moving forward, as the information density and architectural efficiency of smaller models continue to increase, getting high-quality, real-time inference on mobile is going to become trivial.

[ Parent | on HN ]

Re:iPhone 17 Pro Demonstrated Running a 400B LLM by volemo on Monday March 23, 2026 @12:23PM (#47491615)

> moving forward, as the information density and architectural efficiency of smaller models continue to increase

If they continue to increase.

[ Parent | on HN ]

Re:iPhone 17 Pro Demonstrated Running a 400B LLM by vessenes on Monday March 23, 2026 @01:00PM (#47492127)

They will. Either new architectures will come out that give us greater efficiency, or we will hit a point where the main thing we can do is shove more training time onto these weights to get more per byte. Similar thing is already happening organically when it comes to efficient token use; see for instance https://github.com/qlabs-eng/slowrun.

[ Parent | on HN ]

Re:iPhone 17 Pro Demonstrated Running a 400B LLM by simopa on Monday March 23, 2026 @01:20PM (#47492398)

Thanks for the link.

[ Parent | on HN ]

Re:iPhone 17 Pro Demonstrated Running a 400B LLM by simopa on Monday March 23, 2026 @01:50PM (#47492814)

The "if" is fair. But when scaling hits diminishing returns, the field is forced to look at architectures with better capacity-per-parameter tradeoffs. It's happened before, maybe it'll happen again now.

[ Parent | on HN ]

Re:iPhone 17 Pro Demonstrated Running a 400B LLM by anemll on Monday March 23, 2026 @01:49PM (#47492792)

Probably 2x speed for Mac Studio this year if they do double NAND ( or quad?)

[ Parent | on HN ]