Alibaba Unveils Powerful Qwen 3 AI Model Family

Alibaba has launched Qwen 3, a series of advanced AI models designed to compete with leading models from Google and OpenAI. These "hybrid" models offer both fast responses and complex reasoning capabilities.

Open Source Availability and Model Sizes

Many Qwen 3 models are available for download under an open license on Hugging Face and GitHub. These models range from 0.6 billion to 235 billion parameters, reflecting varying levels of problem-solving abilities.

Enhanced Reasoning and Multilingual Support

Qwen 3 models excel at "reasoning," allowing them to fact-check and handle complex problems. They support 119 languages and were trained on a massive dataset of nearly 36 trillion tokens, including text, code, and AI-generated data.

Benchmark Performance and Comparisons

Alibaba claims Qwen 3 significantly outperforms its predecessor, Qwen 2. The largest Qwen 3 model (Qwen-3-235B-A22B) surpasses OpenAI's o3-mini and Google's Gemini 2.5 Pro on certain benchmarks, including Codeforces, AIME, and BFCL. However, this model is not yet publicly available.

The currently available Qwen 3-32B model also demonstrates competitive performance, outperforming OpenAI's o1 on several tests, including LiveBench.

Tool-Calling and Cloud Availability

Qwen 3 boasts strong tool-calling capabilities and excels at following instructions and handling data formats. Beyond open-source downloads, Qwen 3 is accessible through cloud providers like Fireworks AI and Hyperbolic.

Impact on the AI Landscape

Experts believe Qwen 3's open-source nature and competitive performance will significantly impact the AI landscape. It highlights the growing trend of open models rivaling closed-source systems, even amidst restrictions on chip sales to Chinese companies.

“The U.S. is doubling down on restricting sales of chips to China and purchases from China, but models like Qwen 3 that are state-of-the-art and open […] will undoubtedly be used domestically,” Tuhin Srivastava, co-founder and CEO of AI cloud host Baseten, told TechCrunch. “It reflects the reality that businesses are both building their own tools [as well as] buying off the shelf via closed-model companies like Anthropic and OpenAI.”

The rise of models like Qwen 3 increases pressure on American labs to innovate and underscores the global nature of AI development.