Alibaba launched Qwen 2.5-1M: A Strategic Move Against Deep Seek?

Alibaba has rolled out its latest artificial intelligence offering, Qwen 2.5-1M, with claims to have beaten rival solutions such as Deep Seek’s R1 and OpenAI’s ChatGPT across various benchmarks. This new model incorporates a design for handling lengthy inputs that memory-intensive applications demand.

Most unusual, however, was the release of Qwen 2.5-Max on the first day of the Lunar New Year, a time when most Chinese employees were off work celebrating with family. The unforeseeable timing indicates the pressure ramping up against Deep Seek, a Chinese AI startup that has risen to fame within the last three weeks, upsetting not only the global contestants’ chances but attempting to secure Alibaba’s standing in the domestic AI race. 

Also read: India achieves historic space feat with successful SpaDeX Docking Mission

Deep Seek, a barely 20-month-old startup created in Alibaba’s own backyard in Hangzhou, has rattled the U.S. world of tech this week. Meanwhile, Alibaba Cloud has fueled the raging fires with the release of benchmark scores for its own AI that purportedly beats the OpenAI and Anthropic models in some significant dimensions.

The battle for users has prompted cloud giants like Alibaba and Tencent to reduce prices aggressively for the past several months. With Deep Seek joining the fun, the competition has become hotter against the backdrop of half a dozen other fast-rising Chinese AI startups that have gone on to secure unicorn-level funding.

Deep Seek’s rapid success has initiated an all-out war among the Chinese tech companies. Just two days after Deep Seek launched the R1 model, ByteDance, parent company of TikTok, responded with a new AI model of its own, alleging it outperformed OpenAI’s o1, which happens to be Microsoft’s darling investment, on a key benchmark for complex instruction execution.

Also read: Amazon’s Great Freedom Festival Big Discounts on Electronics & Fashion

Deep Seek has gone on record, too, contending that its R1 model is an equal to OpenAI’s systems across different metrics.