Chinese tech giant Alibaba announced the release of its Qwen 2.5 AI model, claiming it surpasses the DeepSeek-V3 model, which has taken the tech world by storm in recent weeks. The launch, which came on the first day of the Lunar New Year – a major holiday in China – has ignited discussions on how the domestic AI market is heating up as Chinese companies fight to establish dominance in the global arena.
The announcement made by Alibaba’s cloud division on January 29 describes the Qwen 2.5-Max model as outperforming DeepSeek-V3 and other leading AI models, including OpenAI’s GPT-4 and Meta’s Llama-3.1-405B. According to Alibaba, the Qwen 2.5-Max’s performance excels across a range of critical benchmarks including complex tasks such as coding and chatbot conversations. Benchmarking results posted on Alibaba’s GitHub page suggest that, while Qwen 2.5 performs comparably to OpenAI’s Claude 3.5-Sonnet, it still holds a significant edge over its competitors in key areas.
Alibaba’s Qwen 2.5 announcement comes at a time when the AI industry, particularly in China, has seen a rapid surge in competition. Just weeks before, DeepSeek, a Chinese AI startup, launched its own DeepSeek-V3 model, followed by the open-source R1 model on January 20. DeepSeek’s aggressive pricing strategy, offering its models at an unprecedented low cost, has disrupted the AI sector globally, and its success has forced other tech companies to respond quickly.
Since DeepSeek’s entry into the AI space, tech stocks in Silicon Valley have taken a hit, and investors have begun questioning the business models of major U.S. companies in the AI sector, which have been heavily reliant on high development costs. In China, DeepSeek’s low-cost approach has put pressure on its competitors, including Alibaba, which is now positioning its Qwen 2.5 as a viable alternative to the startup’s powerful models.
This move by Alibaba, particularly releasing the model during the Lunar New Year when many in China are off work, highlights the growing urgency to address the competition posed by DeepSeek. “The timing of Qwen 2.5’s release is no coincidence. It’s clear that the emergence of DeepSeek has forced Chinese tech giants like Alibaba to accelerate their AI developments,” said Neil Shah, partner and co-founder of Counterpoint Research.
The Price War and Innovation Battle
DeepSeek’s success is also driving a price war within China’s AI sector. DeepSeek-V2, the predecessor to the V3 model, disrupted the market in May 2023 by offering open-source access at a fraction of the cost of its competitors—just 1 yuan (about $0.14) per million tokens processed by the AI. This low-cost strategy forced Alibaba’s cloud division to slash prices by as much as 97% on some of its own models.
This kind of price slashing has caught the attention of global players, including OpenAI and Meta, which are now facing increased pressure to make their own offerings more affordable and accessible to users. ByteDance, the parent company of TikTok, joined the fray in early January by releasing an update to its AI models, claiming that its latest version surpassed OpenAI’s GPT-4o on a critical benchmark test called AIME, which evaluates how well AI models understand and respond to complex instructions.
Despite these intense price cuts, DeepSeek has maintained a narrow user base, especially in comparison to Alibaba, which has hundreds of thousands of employees and an expansive customer base. DeepSeek’s ability to achieve such rapid success on a relatively small budget—$5.6 million for its R1 model—has raised eyebrows, particularly as U.S. companies continue to invest billions in developing similar technologies.
As 2025 unfolds, the global AI market is at a crossroads. The entry of companies like DeepSeek has certainly shaken up the status quo, but with Alibaba’s new Qwen 2.5 model and other domestic players ramping up their efforts, we are likely to see a fierce and evolving landscape. As the competition for AI dominance heats up, both large and small companies will have to find new ways to innovate—whether by cutting costs, improving performance, or striking a balance between the two.
What is clear is that we are witnessing the beginning of a new era in AI, one where domestic and global players alike are pushing the boundaries of what’s possible.
