Qwen 2.5 3B Instruct

Overview

Alibaba's compact 3B Qwen 2.5 instruct model with surprisingly strong MMLU 65.6 and HumanEval 74.4. Locked to non-commercial use under the Qwen Research License.

When to pick this model

• Edge and on-device inference where 2GB VRAM is the budget
• Multilingual prototypes before scaling up to a larger model
• Research and personal projects under the Qwen Research License
• Latency-critical paths where larger models are too slow

VRAM requirements by quantization

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.

Published benchmark scores

Scores published by the model author or aggregated from public leaderboards. Re-measured monthly by our editorial team.

Strengths

• Around 2GB VRAM at Q4 — runs on almost anything
• Multilingual coverage rare at this size
• MMLU 65.6 and HumanEval 74.4 punch above its weight
• 32k context out of the box

Limitations

• Qwen Research License blocks commercial use
• Quality gap vs 7B-and-up is meaningful for non-trivial tasks
• 32k context limits long-document work

Architecture & training

Architecture: Dense transformer · Qwen 2.5 3B · compact multilingual

Training: Full Qwen 2.5 corpus, 3B compressed version.

Quick start

ollama run qwen2.5:3b

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