Alibaba Releases Its Own AI Model, Claiming it Rivals DeepSeek-V3

Chinese tech company Alibaba surprised the Chinese public on Wednesday, the first day of the Lunar New Year, when it released a new version of its Qwen 2.5 artificial intelligence model. Alibaba claims that Qwen 2.5 “surpasses” the highly-acclaimed DeepSeek-V3, according to reports at Reuters.

The timing of release on a holiday-when people are at home with their families- is considered unusal in China and has raised eyebrows, and caused analysts to speculate that either Alibaba feels pressure not to lose ground to DeepSeek, or that it did it to ensure that the release garners maximum attention on the public’s day off.

Alibaba’s cloud unit said in an announcement posted on its official WeChat account, that “Qwen 2.5-Max outperforms … almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B.”

The news comes on the heels of the release of DeepSeek’s AI assistant on January 10, powered by the DeepSeek-V3 model, as well as the Jan. 20 release of its R1 model, which sent the U.S.’s tech hub Silicon Valley reeling and shares of U.S. tech companies, especially Nvidia, tumbling.

The Chinese startup responsible for the latest version of DeepSeek managed to the create the AI model with less resources, lower costs, and ‘rooky’ engineers than ChatGPT, prompting investors to question huge spending plans by leading U.S. AI firms. DeepSeek has also caused an upset in its own domestic market, however, with other Chinese tech firms rushing to release updates to their AI models.

Reuters points out that two days after the release of DeepSeek-R1, TikTok owner ByteDance released an update to its flagship AI model.

ByteDance says it outperforms Microsoft-backed OpenAI’s o1 in tests that measure how well AI models understand and respond to complex instructions.

Source: tovima.com