Qwen

What is Qwen?

Qwen (Tongyi Qianwen) is Alibaba Cloud's AI model family and, next to Llama, the largest open model ecosystem in the world. Dozens of variants from a few billion parameters up to frontier size, many with open weights on Hugging Face. The current flagship is Qwen3.8-Max, generally available since August 3, 2026.

Using it for free

The chat app (web, iOS, Android, macOS) is permanently free with no communicated limits. Alibaba ended the free developer API tier in April 2026; new cloud accounts still get a trial of roughly 70 million tokens (1M per model, 90 days, Singapore endpoint).

API pricing

The range is wide: Qwen-Flash from $0.05 input / $0.40 output per million tokens, Qwen-Plus from $0.40/$1.20, the Qwen3.8-Max flagship at $2/$6. That undercuts Western frontier pricing substantially and competes head-on with DeepSeek for the cheapest-usable-model title.

The real strength: open weights

For companies the interesting part is not the cloud chat but self-hosting: Qwen models run locally or on EU hosting without data flowing to China. The size range makes it easy to match the model to your hardware, from edge minis to large reasoning models. Multilingual performance is a genuine strength.

How it compares

Against ChatGPT and Gemini, Qwen competes on price and openness. Against DeepSeek, Qwen offers the broader model palette while DeepSeek counters with its fully free chat and cheaper flagship. For European data sovereignty without self-hosting, Mistral remains the first pick.

Which model belongs in your stack? Our AI consulting answers that, hosting and GDPR assessment included.