Can Gemma 3 4B run on iPhone 15?
YES — Runs, barely
Grade B · 68/100 · formula estimate
What this means
The checked Q4_K_M version should run, but there is very little memory left for the operating system and other apps. The smaller IQ4_XS version is the safer choice.
Download the IQ4_XS version, which is about 2.3 GB. We expect it to use about 3.3 GB of the roughly 3.9 GB available to a model on this phone.
At the estimated speed, a roughly 300-word answer may take about 40 seconds to finish.
This result is calculated from the phone and model specifications. It has not been measured on this exact phone and setup.
9.2tokens/s
estimated · faster than you read
3.5GB
needed at Q4_K_M
Q4_K_M
quant checked
B
grade · 68/100
needs 3.5 GBusable 3.9 GB
Apple A16 Bionic6 GB RAMMetal
Make it bearable
1
Drop to IQ4_XS (2.3 GB) — it runs comfortably here at ~10 tokens/s. Recommended over Q4_K_M.
2
Drop context to 2K — saves ~0.2 GB of KV cache and a little speed.
3
Close every other app before loading — the 0.4 GB headroom is real; iOS may evict the model otherwise.
4
Expect throttling after ~10 min of sustained generation on a phone chassis.
See how fast it feels
Estimated at 9.2 tokens/s — faster than you read. A ~300-word reply takes about 43 seconds on this phone.
Where the memory goes
| Component | Detail | GB |
|---|---|---|
| Model weights | Q4_K_M GGUF (2.5 GB) + mmap overhead | 2.6 |
| KV cache | 4K context window | 0.3 |
| Runtime | llama.cpp + app overhead | 0.6 |
| Total needed | at Q4_K_M, 4K context | 3.5 |
| Usable on iPhone 15 | 6 GB RAM − iOS reserve (jetsam limit) | 3.9 |
| Headroom | left for the OS and your other apps | 0.4 |
Pick your quant
| Quant | Download | Verdict | Speed |
|---|---|---|---|
| IQ1_S | 1.2 GB | ✓ Runs great | ~19.2 tokens/s |
| Q2_K | 1.7 GB | ✓ Runs great | ~13.6 tokens/s |
| Q3_K_M | 2.1 GB | ✓ Runs great | ~11 tokens/s |
| IQ4_XS BEST HERE | 2.3 GB | ✓ Runs great | ~10 tokens/s |
| Q4_0 | 2.4 GB | ! Runs, barely | ~9.6 tokens/s |
| Q4_K_S | 2.4 GB | ! Runs, barely | ~9.6 tokens/s |
| Q4_K_M | 2.5 GB | ! Runs, barely | ~9.2 tokens/s |
| Q5_K_M | 2.8 GB | ! Runs, barely | ~8.2 tokens/s |
| Q6_K | 3.2 GB | ✕ Won't fit | won't fit |
| Q8_0 | 4.1 GB | ✕ Won't fit | won't fit |
Get it running in 3 steps
1
Install PocketPal from the App Store — free, no account needed.
2
Download "Gemma 3 4B IQ4_XS" in the in-app model hub — 2.3 GB, use Wi-Fi. Search Hugging Face for
unsloth/gemma-3-4b-it-GGUF and pick the IQ4_XS file.3
Set context to 4K and enable Metal — expect ~10 tokens/s on this phone.
Related checks
MORE ON IPHONE 15
GEMMA 3 4B ON OTHER PHONES