text model · gpt-oss · iOS
Can I run gpt-oss 20B on iPhone 17?
No. gpt-oss 20B needs ~13.2 GB even at Q4_K_M, but iPhone 17 only has ~4.5 GB usable.
Needs ~13.2 GB even at Q4_K_M, but only ~4.5 GB is usable.
That figure is at a 4k context and moves about ±15% as context length changes.
The gap is about 8.7 GB: gpt-oss 20B needs roughly 13.2 GB at Q4_K_M and iPhone 17 leaves only about 4.5 GB usable for a model. The lightest tracked hardware that runs gpt-oss 20B is the Nvidia GeForce RTX 4060 Ti (16GB) at 16 GB. See gpt-oss 20B on Nvidia GeForce RTX 4060 Ti (16GB).
- Q4_K_M needed
- ~13.2 GB
- Usable on device
- ~4.5 GB
- Device memory
- 8 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
- 21B (MoE, 3.6B active)
- Q4_K_M size
- 11.28 GB
- Q8_0 size
- 0.86 GB
- Context
- 128k
- Ollama tag
- gpt-oss:20b
- Memory
- 8 GB unified
- Usable for weights
- ~4.5 GB
- Best runtime
- llama.cpp + Metal (via PocketPal or Off Grid app)
What you can run instead
Run gpt-oss 20B on other hardware
FAQ
Can iPhone 17 run gpt-oss 20B?
No. gpt-oss 20B needs ~13.2 GB even at Q4_K_M, but iPhone 17 only has ~4.5 GB usable.
How much memory does gpt-oss 20B need?
iPhone 17 does not have enough memory. At Q4_K_M the weights are ~11.28 GB; with KV cache and runtime overhead, budget ~13.2 GB at a 4k context. It is a Mixture-of-Experts model (21B total / 3.6B active), so all experts must stay in memory; memory tracks total params, not active params.
What is the best tool to run gpt-oss 20B 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/gpt-oss-20b/iphone-17) Sources
-
apple.com · 1 source
-
developer.apple.com · 1 source
-
en.wikipedia.org · 2 sources
-
enclaveai.app · 1 source
-
github.com · 3 sources
-
huggingface.co · 2 sources
-
layla-network.ai · 1 source
-
macrumors.com · 1 source
-
ollama.com · 2 sources
-
privatellm.app · 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.