Alibaba has launched new AI models, including the multimodal Qwen3-Omni, designed to rival OpenAI’s ChatGPT-4, ahead of its annual Apsara Conference in Hangzhou.

Qwen3-Omni is designed to process text, audio, image, and video inputs, producing responses in text and audio.

These releases increase competition with global AI firms, including OpenAI and Google.

Alibaba said Qwen3-Omni supports 119 text languages and 19 spoken languages, and it can generate spoken responses in 10 languages.

Three versions of Qwen3-Omni are now available on open-source platforms such as Hugging Face and GitHub.

The company also released Qwen-Image-Edit-2509, an updated open-source image tool, and Qwen3-TTS-Flash, a proprietary speech model accessible through Alibaba’s cloud platform.

Alibaba claims that variants of Qwen3-Omni outperformed its previous model and international competitors in some benchmark tests, though these results have not been independently verified.

🔗 Source: South China Morning Post

Recent Alibaba developments