DeepSeek news: Chinese tech company Alibaba on Wednesday released a new version of its Qwen 2.5 artificial intelligence model that it claimed surpassed the highly acclaimed DeepSeek-V3, news agency Reuters reported.
The unusual timing of the Qwen 2.5-Max’s release, on the first day of the Lunar New Year when most Chinese people are off work and with their families, points to the pressure Chinese AI startup DeepSeek’s meteoric rise in the past three weeks has placed on not just overseas rivals, but also its domestic competition.
“Qwen 2.5-Max outperforms … almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B,” Alibaba’s cloud unit said in an announcement posted on its official WeChat account, referring to OpenAI and Meta’s most advanced open-source AI models.
Alongside Tencent Holdings Ltd. and Baidu Inc., Alibaba has poured significant resources into its cloud services segment and is engaged in a hot contest to recruit China’s AI developers to use its tools, Bloomberg reported.
DeepSeek, a 20-month-old startup, which was founded in Alibaba’s home city, Hangzhou, has sent shockwaves across tech firms in the US this week. Alibaba Cloud also shared scores that suggest its AI beats OpenAI and Anthropic’s models in certain benchmarks.
Cloud service providers such as Alibaba and Tencent have slashed their pricing in recent months in to win over more users. DeepSeek has already contributed to that price war, alongside a half dozen other promising AI startups in China that have secured funding at unicorn valuations.
DeepSeek R1 rattles tech firms
DeepSeek has been shaking things up in the AI world. Its AI assistant, powered by the DeepSeek-V3 model, launched on January 10, followed by the R1 model on January 20. The impact was immediate—tech stocks took a hit as investors started questioning whether big-budget AI projects in the US were really worth it, given DeepSeek’s reportedly low development and usage costs.
Back home, the startup’s success has kicked off a fierce competition among Chinese tech giants. Just two days after the R1 model dropped, TikTok’s parent company ByteDance fired back with an update to its own flagship AI model, claiming it outperformed OpenAI’s o1—Microsoft’s AI bet—in a key benchmark test that measures how well AI understands and responds to complex instructions.
DeepSeek, of course, had already made a similar claim, saying its R1 model could go head-to-head with OpenAI’s o1 on multiple performance benchmarks.
(With inputs from Reuters, Bloomberg)