text model · Granite · Android
Can I run Granite 3.1 2B on Generic Android Phone (12GB RAM)?
Yes. Granite 3.1 2B runs on Generic Android Phone (12GB RAM) at Q4_K_M (~2.8 GB of ~8.5 GB usable).
Runs at Q4_K_M using ~2.8 GB of ~8.5 GB usable. You have room for FP16 for higher quality.
That figure is at a 4k context and moves about ±15% as context length changes. Generic Android Phone (12GB RAM) leaves ~5.7 GB of headroom, room to step up to FP16 for higher quality.
- Q4_K_M needed
- ~2.8 GB
- Usable on device
- ~8.5 GB
- Device memory
- 12 GB
- Best quant
- Q4_K_M
Which quant fits
How to run it
On Android use PocketPal AI (Polished app, download GGUF and run offline.).
- Parameters
- 2.53B
- Q4_K_M size
- 1.55 GB
- Q8_0 size
- 2.69 GB
- Context
- 128k
- Ollama tag
- granite3.1-dense:2b
- Memory
- 12 GB ram
- Usable for weights
- ~8.5 GB
- Best runtime
- llama.cpp (PocketPal) or MLC-LLM
You could also run
Run Granite 3.1 2B on other hardware
FAQ
Can Generic Android Phone (12GB RAM) run Granite 3.1 2B?
Yes. Granite 3.1 2B runs on Generic Android Phone (12GB RAM) at Q4_K_M (~2.8 GB of ~8.5 GB usable).
How much memory does Granite 3.1 2B need?
Generic Android Phone (12GB RAM) has room to spare. At Q4_K_M the weights are ~1.55 GB; with KV cache and runtime overhead, budget ~2.8 GB at a 4k context.
What is the best tool to run Granite 3.1 2B on Android?
On Android, PocketPal AI (Polished app, download GGUF and run offline.) is the go-to option. NPU acceleration is limited and chip-specific; most apps run on CPU. Expect 1B-4B class.
Embed this
A live badge for your model card or README, updated as the data is. [](https://localmodel.run/can-i-run/granite-3.1-2b/android-generic-12gb) Sources
- community.ibm.com
- github.com/a-ghorbani
- github.com/google-ai-edge
- github.com/mlc-ai
- github.com/shubham0204
- github.com/Vali-98
- huggingface.co/bartowski/granite-3.1-2b-instruct-GGUF
- huggingface.co/bartowski/Meta-Llama-3.1-8B-Instruct-GGUF
- huggingface.co/ibm-granite
- layla-network.ai
- ollama.com/library/granite3.1-dense:2b
- ollama.com/library/llama3.1:8b
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.