text model · Mistral · iOS
Can I run Ministral 3 14B on iPhone 17 Pro?
No. Ministral 3 14B needs ~9.4 GB even at Q4_K_M, but iPhone 17 Pro only has ~8 GB usable.
Needs ~9.4 GB even at Q4_K_M, but only ~8 GB is usable.
That figure is at a 4k context and moves about ±15% as context length changes.
The gap is about 1.4 GB: Ministral 3 14B needs roughly 9.4 GB at Q4_K_M and iPhone 17 Pro leaves only about 8 GB usable for a model. The lightest tracked hardware that runs Ministral 3 14B is the Nvidia GeForce RTX 3060 (12GB) at 12 GB. See Ministral 3 14B on Nvidia GeForce RTX 3060 (12GB).
- Q4_K_M needed
- ~9.4 GB
- Usable on device
- ~8 GB
- Device memory
- 12 GB
Which quant fits
How to run it
On iOS use Apple Foundation Models (Built into iOS 26, ~3B on-device model, zero download, fully private.).
- Parameters
- 14B
- Q4_K_M size
- 7.67 GB
- Q8_0 size
- 13.37 GB
- Context
- 256k
- Ollama tag
- ministral-3:14b
- Memory
- 12 GB unified
- Usable for weights
- ~8 GB
- Power draw
- ~12 W
- Best runtime
- llama.cpp + Metal (via PocketPal or Off Grid app)
What you can run instead
Run Ministral 3 14B on other hardware
FAQ
Can iPhone 17 Pro run Ministral 3 14B?
No. Ministral 3 14B needs ~9.4 GB even at Q4_K_M, but iPhone 17 Pro only has ~8 GB usable.
How much memory does Ministral 3 14B need?
iPhone 17 Pro does not have enough memory. At Q4_K_M the weights are ~7.67 GB; with KV cache and runtime overhead, budget ~9.4 GB at a 4k context.
What is the best tool to run Ministral 3 14B on iOS?
On iPhone and iPad, Apple Foundation Models (Built into iOS 26, ~3B on-device model, zero download, fully private.) is the standard choice. Phones realistically run 1B-4B class models. Anything larger thermally throttles or OOMs.
Embed this
A live badge for your model card or README, updated as the data is. [](https://localmodel.run/can-i-run/ministral-3-14b/iphone-17-pro) Sources
-
apple.com · 1 source
-
developer.apple.com · 1 source
-
en.wikipedia.org · 1 source
-
enclaveai.app · 1 source
-
github.com · 3 sources
-
layla-network.ai · 1 source
-
macrumors.com · 1 source
-
ollama.com · 1 source
-
privatellm.app · 1 source
-
wccftech.com · 1 source
Weights are measured from GGUF files; KV cache and overhead are computed, so totals can vary ~15% with context and runtime. Any tok/s is a bandwidth estimate. See methodology.