README.md

---
title: Dispatch AI — Mobile AI Speed Test
emoji: ⚡
colorFrom: green
colorTo: blue
sdk: gradio
sdk_version: "4.44.1"
app_file: app.py
pinned: true
license: other
tags:
  - benchmark
  - mobile
  - ai
  - speed-test
  - dispatch-ai
---

# Dispatch AI — Mobile AI Speed Test ⚡

Like Speedtest, but for AI. Pick a phone model → see benchmark results for popular LLMs. Compare phones side-by-side.

Built by **Dispatch AI (FZE)** — UAE · License 10818.

## Features

- 📱 **Single Phone Test** — pick a phone, see generation/prompt speeds, RAM, load time for 9+ models
- ⚔️ **Compare Two Phones** — side-by-side comparison with speed difference and percentage
- 📊 **6 phones** in the farm:
  - Samsung S20 FE (SD865) — 40 devices
  - Samsung S22 (SD8 Gen1) — 12 devices
  - Samsung A54 (Exynos 1380) — 8 devices
  - Pixel 7 (Tensor G2) — 6 devices
  - OnePlus 11 (SD8 Gen2) — 4 devices
  - Xiaomi Redmi Note 12 (SD685) — 10 devices
- 🌙 **Dark theme** (#0A0F1A bg, #1FE0E6 cyan accent)

## Methodology

- **Backend:** `llama.cpp` (via `llamafile`), Q4_K_M quants, 4 CPU threads, FP16 offload
- **Prompts:** Fixed 512-token prompt; 256 generation tokens; batch size 512
- **Aggregation:** Median across all devices of each model after 5-run warm-up
- **Environment:** Air-conditioned room at 22°C; screens off; batteries 80–100%

## Models Tested

Qwen2.5-0.5B, Qwen2.5-1.5B, Llama-3.2-1B, Llama-3.2-3B, Gemma-2-2B, Phi-3.5-mini, SmolLM2-1.7B, SmolLM2-135M, TinyLlama-1.1B (all Q4_K_M).

## License

© 2026 Dispatch AI FZE — Sharjah, UAE · License 10818. All rights reserved.