Sat Mar 08 17:50:00 UTC 2025: ## Alibaba’s New AI Model Outperforms OpenAI and DeepSeek, Sending Shares Soaring

**HONG KONG –** Chinese tech giant Alibaba has unveiled its latest AI reasoning model, QwQ-32B, claiming it surpasses rival models from OpenAI and DeepSeek in performance. The announcement sent Alibaba’s Hong Kong-listed shares up 8%, boosting the Hang Seng’s China Enterprises Index.

Alibaba boasts that QwQ-32B’s performance is “exceptional,” outperforming OpenAI’s o1-mini and rivaling DeepSeek’s R1, a leading open-source model, despite having significantly fewer parameters (32 billion compared to R1’s 671 billion). Fewer parameters indicate greater efficiency in training the model.

This launch follows the unveiling of Manus, a “general AI agent” by Chinese company Monica, capable of complex multi-step tasks. Alibaba’s aggressive push into AI comes amidst increased government support for the sector in China and follows a pledge by Alibaba to invest at least $52.4 billion in AI and cloud computing infrastructure over the next three years.

The rapid advancements in Chinese AI technology, highlighted by both Alibaba’s and DeepSeek’s recent successes, are bolstering investor confidence in the face of intensifying US-China tech rivalry. The Hang Seng China Enterprises Index has seen a significant 30% surge since January. Alibaba’s previous AI models, including Tongyi Qianwen and Qwen 2.5 Max, have also demonstrated strong performance against international competitors. The company’s latest achievement underscores China’s growing prowess in the rapidly evolving field of artificial intelligence.

Read More