Qwen 2.5 Coder 1.5B Instruct

Overview

Alibaba's smallest Qwen 2.5 Coder at 1.5B parameters under Apache 2.0, covering 92 programming languages. HumanEval 70.7 makes it a serious on-device completion model.

When to pick this model

• Local inline completion in IDE plugins
• Edge devices and laptops without dedicated GPUs
• Latency-critical code suggestions where 7B is too slow
• Fallback model when bigger coders are unavailable

VRAM requirements by quantization

Quantization	VRAM required
Q4_K_M (recommended)	1 GB
Q5_K_M	1.2 GB
Q8_0	2 GB
FP16 (no quantization)	3 GB

VRAM figures include model weights plus a typical 8k KV cache and ~600 MB runtime overhead (Ollama / llama.cpp baseline). Add headroom for higher context lengths.

Strengths

• Around 1GB VRAM at Q4 — runs nearly anywhere
• Strong inline completion for a 1.5B model
• Apache 2.0 license
• 92 programming languages covered

Limitations

• 1.5B caps code quality — not for complex generation
• 32k context only
• Outclassed on harder tasks by 7B+ coders

Architecture & training

Architecture: Dense · 1.5B · Qwen 2.5 Coder · compact code-specialized

Training: 1.5B params, code corpus across 92 languages, ideal for lightweight completion.

Quick start

ollama run qwen2.5-coder:1.5b

Or use the open-source MCP server to query this model from Claude Desktop, Cursor, or any MCP-compatible client.