Alibaba has introduced Qwen3, a family of AI models that reportedly match or surpass the capabilities of models from Google and OpenAI. These models, ranging from 0.6 billion to 235 billion parameters, are available under an open license on platforms like Hugging Face and GitHub. Qwen3 models are described as ‘hybrid’ because they can either reason through complex problems or provide quick answers to simpler queries. Some models use a mixture of experts (MoE) architecture for computational efficiency. Qwen3 supports 119 languages and was trained on nearly 36 trillion tokens. While the largest model, Qwen-3-235B-A22B, outperforms some top models on specific benchmarks, it is not yet publicly available. The largest public model, Qwen3-32B, remains competitive with other leading models.
— new from TechCrunch