Qwen Models — Hardware Requirements

21 Qwen models from deepcogito and the community, from the smallest that runs in 0.8 GB of VRAM up to 72.3B parameters. Every row links to full quantization tables and GPU compatibility.

All Qwen Models by Size

ModelParamsContext
SpatialLM1.1 Qwen 0.5B604M33K
Qwen1.5 0.5B Chat620M33K
Nemotron Research Reasoning Qwen 1.5B1.8B131K
Qwen 1 8B1.8B8K
Qwen1.5 1.8B1.8B33K
Qwen1.5 MoE A2.7B Chat2.7B33K
Qwen35 4B Soyuz Merged4B262K
CyberSecQwen 4B4.0B262K
CodeQwen1.5 7B7.3B66K
Qwen1.5 7B Chat7.7B33K
Qwen1.5 7B7.7B33K
Qwen 7B7.7B33K
Qwen Marketing8.2B
Qwen1.5 14B Chat14.2B33K
Qwen 14B Chat14.2B8K
Qwen1.5 14B14.2B33K
Qwen 14B14.2B8K
Qwen1.5 MoE A2.7B14.3B8K
Cogito V1 Preview Qwen 32B32B131K
XiYanSQL QwenCoder 32B 250432B33K
Qwen1.5 32B Chat32.5B33K
Qwen1.5 32B32.5B33K
Qwen1.5 72B Chat72.3B33K

How Qwen Compares — Benchmark Rating

Qwen 14B is the highest-rated Qwen model with an overall benchmark rating of 56.6/100 — #20 among 75 open models. The top proprietary model, GPT 5.5, scores 88.8. Click a model to see its full benchmark breakdown.

GPT 5.5 · proprietary88.8
Claude Opus 4.7 · proprietary87.6
Claude Fable 5 · proprietary86.6
GPT 5.4 · proprietary86.6
Claude Opus 4.8 · proprietary84.4
Composite of normalized public benchmark scores (methodology) · Qwen · other models

Frequently Asked Questions

How much VRAM do I need to run a Qwen model?
The smallest Qwen model, Qwen1.5 0.5B Chat, runs from 0.8 GB of VRAM at an aggressive quantization. Larger family members need proportionally more — see the table above for every model.
Which Qwen models can I run on a 16 GB GPU?
21 of 23 Qwen models fit in 16 GB of VRAM at some quantization, including Cogito V1 Preview Qwen 32B, Qwen1.5 0.5B Chat, Qwen1.5 14B Chat.
What is the most popular Qwen model to run locally?
Cogito V1 Preview Qwen 32B is the most downloaded Qwen model in local-friendly quantized formats. It runs from 10.4 GB of VRAM.
How do Qwen models score on benchmarks?
Qwen 14B leads the family with an overall benchmark rating of 56.6/100, ranking #20 among 75 open models, while the top proprietary model, GPT 5.5, scores 88.8. See the comparison chart above for the full standings.