Skip to content
localmodel.run

audio model · kokoro · iOS

Can I run Kokoro-82M on iPhone 15 Pro?

Compatibility verdict memory check
Yes, it runs heavy offloading

Yes. Kokoro-82M runs on iPhone 15 Pro at fp32 (~1 GB of ~4.5 GB usable).

Needs ~1 GB Device usable ~4.5 GB

Runs at fp32 using ~1 GB of ~4.5 GB usable.

Peak memory
~1 GB
Usable on device
~4.5 GB
Device memory
8 GB
Quant
fp32
Share on X Share on Reddit

How to run it

Use kokoro (Python) or ONNX Runtime at fp32. It is light enough to run on CPU; a GPU just makes it faster.

Model kokoro
Type
Text to speech
Parameters
82M
Peak memory
~1 GB at fp32
License
Apache-2.0
Full Kokoro-82M requirements →
Device iOS
Memory
8 GB unified
Usable for weights
~4.5 GB
Power draw
~14 W
Best runtime
llama.cpp + Metal (via PocketPal or Off Grid app)
Best models for iPhone 15 Pro →

You could also run

Run Kokoro-82M on other hardware

Nvidia GeForce RTX 2060 (6GB)Apple M1 (8GB)8GB RAM Laptop (CPU/iGPU only)iPhone 16iPhone 16 ProGeneric Android Phone (8GB RAM)

FAQ

Can iPhone 15 Pro run Kokoro-82M?

Yes. Kokoro-82M runs on iPhone 15 Pro at fp32 (~1 GB of ~4.5 GB usable).

How much memory does Kokoro-82M need?

iPhone 15 Pro has room to spare. At fp32 the realistic peak is ~1 GB of memory.

What do I use to run Kokoro-82M locally?

Kokoro-82M runs in kokoro (Python) or ONNX Runtime (among others). It runs on CPU, so no GPU is required.

Sources

VRAM figures are sourced peak-usage anchors at the noted quant; catalog updated 2026-10-05. See methodology.