Alibaba Launches Advanced AI Model, Claims Superiority Over DeepSeek-V3

On the first day of the Lunar New Year, a typically quiet time for China’s stock market, Alibaba made a significant announcement by unveiling a new iteration of its AI model, Qwen 2.5. This release came at an unusual time when most of the country is celebrating, suggesting the urgency Alibaba felt in responding to the recent advancements by DeepSeek, a burgeoning Chinese AI startup.

Alibaba’s cloud unit announced that the Qwen 2.5-Max model outperforms several leading AI models, including DeepSeek-V3, GPT-4o by OpenAI, and Llama-3.1-405B by Meta, across a broad spectrum of benchmarks. This bold claim was made through an official WeChat post, highlighting the model’s capabilities in a competitive landscape where AI technology is advancing rapidly.

The meteoric rise of DeepSeek has not only put pressure on international tech giants but also on domestic competitors like Alibaba. DeepSeek’s strategy has been to offer AI models at significantly lower costs, with their DeepSeek-V2 model, for instance, being open-source and priced at just 1 yuan ($0.14) per million tokens. This pricing strategy sparked a price war in the Chinese AI market, leading Alibaba to slash prices on its models by up to 97%.

Liang Wenfeng, DeepSeek’s founder, has articulated a vision centered around achieving Artificial General Intelligence (AGI), a concept where AI can outperform humans in economically valuable tasks. Unlike traditional tech giants with large, structured teams, DeepSeek operates more like an academic lab, with a lean team of young, highly educated professionals from China’s top universities. This approach contrasts with the resource-heavy methods of larger tech corporations, suggesting a new paradigm in AI development.

The success of DeepSeek has prompted a rush among other Chinese tech companies to enhance their AI technologies. Notably, ByteDance, the company behind TikTok, quickly followed DeepSeek’s releases with an update to its AI model, claiming superiority over Microsoft-backed OpenAI’s latest offerings in certain benchmark tests.

Alibaba’s response to this competitive pressure with the Qwen 2.5-Max release indicates a strategic pivot to keep pace or even set the pace in the AI race, especially during a time when the tech community was least expecting it. This move not only showcases Alibaba’s commitment to AI innovation but also underscores the fierce competition within the Chinese tech sector to lead in the global AI arena.

This development is part of a broader narrative where Chinese tech firms are increasingly seen as formidable players in AI, challenging established tech giants worldwide and altering the dynamics of innovation and market competition.