Run Gemma 4 on iPhone 15 with Google AI Edge Gallery

iPhone 15 with the A16 Bionic chip and 6 GB of RAM is the entry-level choice for on-device Gemma. The smaller memory budget rules out the larger Gemma 4 E4B model, but Gemma 4 E2B runs fine through Google AI Edge Gallery at an estimated 15 tokens per second. That is slower than 8 GB iPhones but still useful for chat, translation, and image Q&A on the go. Apple Intelligence is not supported here, so Edge Gallery is your only option for serious on-device AI on this phone.

Apple A16 Bionic6 GB RAM2023

Quick answer

On the iPhone 15 (Apple A16 Bionic, 6GB RAM), run Google Gemma 4 E2B at ~15 tok/s (est.) via Google AI Edge Gallery. Gemma 4 E2B is ~2.5 GB on disk and uses 1-1.5 GB in memory, fully offline, no account required.

→ Install Google AI Edge Gallery (iOS 17+), then download Gemma 4.

TOP PICK

Gemma 4 E2B

SPEED

~15 tok/s (est.)

DEVICE RAM

6 GB

Speed is a ModelFit estimate from chip generation, not a measured benchmark.

Verdict

Recommended: Gemma 4 E2B

iPhone 15 supports Gemma 4 E2B comfortably. The 6 GB RAM ceiling means E4B is off the table. That is fine, E2B is plenty for daily use.

Hardware Profile

DEVICE

iPhone 15

CHIP

Apple A16 Bionic

RAM

6 GB

NEURAL ENGINE

16-core

Gemma 4 Performance on iPhone 15

Model	Download	RAM in Use	Speed	Source
Gemma 4 E2BTop Pick	~2.5 GB	1-1.5 GB	~15 tok/s	Estimated
Gemma 4 E4B	~5 GB	2-3 GB	Not recommended	N/A

Speeds via Google AI Edge Gallery on iOS 17+. "Measured" numbers come from real-world Hacker News user reports; "Estimated" numbers are interpolations from chip generation. Both Gemma 4 variants use int4 quantization-aware training.

Best for

iPhone 15 owners who do not have Apple Intelligence
Casual chat, writing, and translation
Offline use during travel
Privacy-sensitive personal queries

Watch outs

Gemma 4 E4B is not recommended. RAM headroom too tight
Older A16 chip is slower than A17 Pro and A18 generations
Apple Intelligence is not supported on iPhone 15 (non-Pro)

Setup Guide

Step-by-step install for Google AI Edge Gallery on iPhone 15, plus full benchmarks and the privacy details.

Read the full guide

Frequently Asked Questions

Can iPhone 15 run Gemma 4?

Yes, but only the smaller E2B variant. Gemma 4 E2B uses 1-1.5 GB of RAM in active inference, which fits comfortably on iPhone 15's 6 GB. The larger E4B model needs 2-3 GB and is too tight on 6 GB devices.

How fast is Gemma 4 on iPhone 15?

Estimated at roughly 15 tokens per second for Gemma 4 E2B. That is the slowest of the supported iPhones but still fast enough for conversational use. The A16 Bionic chip is two generations behind current models, so newer iPhones will be noticeably faster.

Why does iPhone 15 only have 6 GB of RAM?

Apple reserved 8 GB RAM for the iPhone 15 Pro and Pro Max in 2023, mainly because Apple Intelligence required it. The standard iPhone 15 was specced for the older OS feature set. That is why it cannot run Apple Intelligence and why Gemma 4 E4B is off the table.

Should I upgrade from iPhone 15 for better local AI?

Only if AI is your primary use case. iPhone 16 doubles the Gemma E2B performance and adds Apple Intelligence support thanks to 8 GB of RAM. If you mostly use other apps and just want to try Gemma occasionally, iPhone 15 with E2B works fine for casual use.