Alibaba releases AI technology challenging DeepSeek on Chinese New Year's Day
Alibaba launches a new AI model to compete with DeepSeek on Chinese New Year's Day.
TroibNews 
On Wednesday, Chinese tech giant Alibaba unveiled its Qwen 2.5 artificial intelligence model, asserting that it exceeds the much-lauded DeepSeek-V3. The new model, named Qwen 2.5-Max, made its debut on the first day of the Chinese New Year. Additionally, the company open-sourced its visual model, Qwen 2.5-VL, the previous day. On the same day, Alibaba also provided cloud computing support for the live broadcast of the Spring Festival Gala, a popular annual variety show in the country that garners billions of viewers and features a mix of music, dance, opera, martial arts, and comedy.
"Qwen 2.5-Max outperforms ... almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B," stated Alibaba's cloud unit in an announcement on its official WeChat account, highlighting a comparison with OpenAI and Meta's leading open-source AI models.
The January 10 launch of DeepSeek's AI assistant, which runs on the DeepSeek-V3 model, has sent shockwaves through Silicon Valley, leading to a downturn in tech stocks. Investors have begun to scrutinize the substantial spending plans of major AI firms in the United States, partly due to DeepSeek's reportedly low development and usage costs.
DeepSeek's rapid ascent has ignited a race among domestic rivals to enhance their AI offerings. Just two days following the introduction of DeepSeek-R1, ByteDance, the parent company of TikTok, rolled out an update to its flagship AI model, claiming it outperformed Microsoft-supported OpenAI's o1 in AIME, a benchmark that evaluates how effectively AI models comprehend and react to intricate instructions.
This latest development mirrors DeepSeek's assertion that its R1 model competes equally with OpenAI's o1 in various performance metrics.
Navid Kalantari for TROIB News
"Qwen 2.5-Max outperforms ... almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B," stated Alibaba's cloud unit in an announcement on its official WeChat account, highlighting a comparison with OpenAI and Meta's leading open-source AI models.
The January 10 launch of DeepSeek's AI assistant, which runs on the DeepSeek-V3 model, has sent shockwaves through Silicon Valley, leading to a downturn in tech stocks. Investors have begun to scrutinize the substantial spending plans of major AI firms in the United States, partly due to DeepSeek's reportedly low development and usage costs.
DeepSeek's rapid ascent has ignited a race among domestic rivals to enhance their AI offerings. Just two days following the introduction of DeepSeek-R1, ByteDance, the parent company of TikTok, rolled out an update to its flagship AI model, claiming it outperformed Microsoft-supported OpenAI's o1 in AIME, a benchmark that evaluates how effectively AI models comprehend and react to intricate instructions.
This latest development mirrors DeepSeek's assertion that its R1 model competes equally with OpenAI's o1 in various performance metrics.
Navid Kalantari for TROIB News
