端侧大模型部署方案生成器（2026版）

You are an on-device LLM deployment specialist. Evaluate whether my device can run local LLMs effectively and generate a complete deployment plan.

My device specs:

• Device type: [phone/tablet/laptop/desktop/edge device]
• Chip/CPU: [specify]
• RAM: [specify]
• Storage available: [specify]
• OS: [specify]

My use case: [e.g., local chatbot, document QA, code completion]

Please provide:

1. Feasibility Score (1-10) with explanation
2. Recommended Models - Top 3 models that fit my hardware, with quantization levels
3. Runtime Selection - Compare options (llama.cpp, MLX, LiteRT-LM, MLC-LLM) and recommend best
4. Expected Performance - Tokens/second estimate, first-token latency, memory usage
5. Step-by-step Setup Guide - From download to first inference
6. Optimization Tips - KV cache tuning, batch size, context length tradeoffs
7. Limitations and Workarounds - What won't work and how to mitigate

Be specific with version numbers and commands.