Qwen
In plain wordsA family of AI language models from the Chinese company Alibaba, many free to download, which by 2025 had become the most popular starting point for other people's open AI models.
Alibaba Cloud's series of large language modelslarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code., issued since 2023 in a wide range of sizes; many are released as open-weightopen-weight model, n.: An AI model whose trained internal numbers are published for anyone to download and run on their own computers, even if the data and code used to make it stay private. models under the permissive Apache 2.0 license, while larger proprietary 'Max' models are offered only as a service on Alibaba's cloud. The family spans text, coding (Qwen-Coder), vision–language (Qwen-VL), audio and 'omni' multimodal models, and the reasoning model QwQ (November 2024). Qwen3 (April 2025) introduced hybrid models that can switch between quick answers and step-by-step thinkingextended thinking, n.: A setting in some AI models, notably Anthropic's Claude, that lets the model write out private step-by-step reasoning before its final answer, improving results on hard problems., in dense and mixture-of-expertsmixture of experts, n.: A way to build very large AI models cheaply: the model contains many specialist sub-networks, and for each word only a few are switched on, picked by a small dispatcher network. versions of up to 235 billion parameters; Qwen3.5 followed in February 2026. Strong results at small and medium sizes and a rapid release cadence made Qwen, by 2025, the most common base for derivative models on Hugging Face, ahead of LlamaLlama, prop. n.: A family of AI language models from Meta, first released in 2023, whose trained settings can be downloaded and run by anyone, making them the foundation for thousands of other AI projects., and in 2026 Alibaba said Qwen models had been downloaded more than 3 billion times.
Half the open models on the leaderboard turned out to be Qwen fine-tunes.
Origin
Alibaba launched the Tongyi Qianwen chatbot in beta in April 2023 and released its first open weights, Qwen-7B, on 3 August 2023. 'Qwen' is short for Qianwen, roughly 'a thousand questions', from the Chinese name.
Usage note
Like other Chinese models, Qwen is trained to comply with Chinese content rules, which shapes its answers on politically sensitive topics. Its open releases, with DeepSeek-R1DeepSeek-R1, prop. n.: A free, downloadable AI reasoning model released by the Chinese company DeepSeek in January 2025. It matched leading American models at math and coding for a fraction of the reported cost, shaking the industry., made Chinese laboratories leaders in open-weight AI by 2025.
Related
- open-weight modelopen-weight model, n.: An AI model whose trained internal numbers are published for anyone to download and run on their own computers, even if the data and code used to make it stay private.
- LlamaLlama, prop. n.: A family of AI language models from Meta, first released in 2023, whose trained settings can be downloaded and run by anyone, making them the foundation for thousands of other AI projects.
- DeepSeek-R1DeepSeek-R1, prop. n.: A free, downloadable AI reasoning model released by the Chinese company DeepSeek in January 2025. It matched leading American models at math and coding for a fraction of the reported cost, shaking the industry.
- large language modellarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code.
- mixture of expertsmixture of experts, n.: A way to build very large AI models cheaply: the model contains many specialist sub-networks, and for each word only a few are switched on, picked by a small dispatcher network.
- DeepSeek momentDeepSeek moment, n.: The shock of January 2025, when a little-known Chinese company released a free AI model nearly as good as the best American ones at a fraction of the reported cost, briefly wiping hundreds of billions off chip stocks.
Sources
- Qwen Team (2025). Qwen3 Technical Report. arXiv:2505.09388.
Mentioned in
LlamaLlama, prop. n.: A family of AI language models from Meta, first released in 2023, whose trained settings can be downloaded and run by anyone, making them the foundation for thousands of other AI projects., AlibabaAlibaba, prop. n.: A Chinese e-commerce and cloud-computing giant whose Qwen family of AI models, mostly free to download, is among the most widely used in the world., DeepSeek momentDeepSeek moment, n.: The shock of January 2025, when a little-known Chinese company released a free AI model nearly as good as the best American ones at a fraction of the reported cost, briefly wiping hundreds of billions off chip stocks., DeepSeek-R1DeepSeek-R1, prop. n.: A free, downloadable AI reasoning model released by the Chinese company DeepSeek in January 2025. It matched leading American models at math and coding for a fraction of the reported cost, shaking the industry., Moonshot AIMoonshot AI, prop. n.: A Beijing artificial intelligence company, founded in 2023, that makes the Kimi chatbot and very large models that anyone can download., Zhipu AIZhipu AI, prop. n.: A Chinese artificial intelligence company that grew out of Tsinghua University and makes the GLM family of language models. Outside China it uses the name Z.ai.