Qwen2.5-1.5B-Instruct-mobile-int4
✅ Verified on real phone hardware — Snapdragon 865, June 2026.
Phone Benchmark (Samsung S20 FE, Snapdragon 865)
| Metric | Value |
|---|---|
| Phone Speed | 16.1 tokens/sec |
| CPU Speed | 7.7 tokens/sec |
| File Size | 1066 MB |
| Chat Format | chatml |
| Test Output | "Paris" ✅ (correct) |
Usage
Python (llama-cpp-python)
from llama_cpp import Llama
llm = Llama(model_path="model.gguf", chat_format="chatml", n_ctx=512, n_threads=4, verbose=False)
response = llm.create_chat_completion(
messages=[{"role": "user", "content": "What is the capital of France?"}],
max_tokens=50,
)
print(response["choices"][0]["message"]["content"])
dispatchAI SDK
from dispatchai import load_model
model = load_model("Qwen2.5-1.5B-Instruct-mobile-int4", backend="gguf")
print(model.chat("What is the capital of France?"))
On Android (via ADB)
hf download dispatchAI/Qwen2.5-1.5B-Instruct-mobile-int4 model.gguf
MSYS_NO_PATHCONV=1 adb push model.gguf /data/local/tmp/
MSYS_NO_PATHCONV=1 adb shell "cd /data/local/tmp && LD_LIBRARY_PATH=/data/local/tmp ./llama-cli -m model.gguf -p 'Hello' -n 30 -t 4 -st"
Model Details
| Attribute | Value |
|---|---|
| Base Model | unknown |
| File Size | 1066 MB |
| Format | GGUF |
| Chat Format | chatml |
| License | apache-2.0 |
About dispatchAI
dispatchAI — Small. Mobile. Free. UAE-built.