Complete directory of 51 Qwen AI models from Alibaba Cloud, ranked by composite score. Compare every Qwen model including Qwen 3, Qwen 2.5, QwQ, and more by pricing, capabilities, context window, and performance. Data refreshed hourly.
Total Models
51
Average Score
50
Free Models
6
With Reasoning
23
Every Qwen model sorted by composite score. Click any row for detailed benchmarks and analysis.
Qwen API pricing across all models. Prices are per 1 million tokens.
How many Qwen models support each capability.
Qwen models are released with open weights under permissive licenses. Teams can self-host, fine-tune on proprietary data, and deploy without vendor lock-in. The Qwen family has become one of the most popular open-source model series, with active community support for quantizations, adapters, and tooling.
Developed by Alibaba Cloud, Qwen models excel at multilingual tasks with particularly strong Chinese and English performance. The training data spans dozens of languages, making Qwen an ideal choice for international applications, translation pipelines, and cross-lingual retrieval tasks.
Qwen models are available through OpenRouter and other API providers at highly competitive per-token rates. Several Qwen variants are offered completely free, and even the largest models undercut many proprietary alternatives, making them ideal for high-volume production workloads.
With open weights, Qwen models can be deployed on your own infrastructure using vLLM, TGI, or Ollama. Self-hosting eliminates per-token API costs, keeps data fully private, and allows custom optimization for your specific hardware. The Qwen family includes models at every size tier from 4B to 235B+ parameters.
40 of 51 Qwen models in our database are marked as open source. Alibaba Cloud has been one of the most prolific contributors to the open-source AI ecosystem, releasing model weights under permissive licenses that allow commercial use, fine-tuning, and redistribution.
The Qwen family spans a wide range of model sizes and specializations. Qwen 3 models represent the latest generation with mixture-of-experts architectures and thinking capabilities, while Qwen 2.5 remains a proven workhorse for coding and general-purpose tasks. The QwQ series focuses on reasoning and chain-of-thought problem solving.
For teams evaluating Qwen vs other open-source models, the breadth of the Qwen lineup is a major advantage. From lightweight 4B-parameter models for edge deployment to 235B+ parameter flagships, Qwen provides options at every scale. All open-weight models can be self-hosted with tools like vLLM, Ollama, or text-generation-inference.
Dive deeper into AI model rankings, pricing breakdowns, and head-to-head comparisons across all providers.