Alibaba’s AI Model Dominates Global Crypto Trading Test, Surpassing GPT-5
According to a report, Alibaba Group’s artificial intelligence model, Qwen3-Max, has surpassed five other prominent AI systems in a real-market cryptocurrency trading experiment carried out by the US research firm Nof1. In the initial round of the Alpha Arena competition, which wrapped up on Tuesday, Alibaba Cloud’s Qwen3-Max recorded a 22.32 percent return on a starting investment of $10,000 within a mere two weeks. Only two of the six competing models achieved profitability. China’s DeepSeek V3.1 Chat achieved a notable second place with a 4.89 percent gain, in contrast to the four US-developed models—OpenAI, Anthropic, Google DeepMind, and Elon Musk’s xAI—which all experienced significant losses. According to the report, OpenAI’s GPT-5 performed the poorest, incurring a loss of 62.66 per cent of its investment. Nof1 cautioned in a blog post that the early results “may be the result of luck,” noting that future rounds would include “more statistical rigour.” The Alpha Arena experiment restricts AI models to quantitative market data exclusively, excluding news or external information, which prompts inquiries regarding the potential effectiveness of these systems in real-world investing scenarios.
“If nothing else, we hope Alpha Arena highlights the power of evaluating AI in more consequential, realistic environments,” Nof1 stated. “We believe this is the fastest path to uncovering critical gaps and insights that move frontier AI forward.” Interestingly, Alibaba’s Qwen3-Max was the sole model lacking reasoning capabilities, which indicates it did not engage in step-by-step thinking prior to executing trades. In the face of increasing competition in artificial intelligence, Alibaba is making strides in its chip development initiatives. An August report states that the company has developed a new AI chip, which is claimed to be more versatile than earlier versions. In contrast to its predecessor, which was produced by Taiwan Semiconductor Manufacturing Company, the new model has been wholly manufactured in China and remains in the testing phase. It is crafted to manage a broader spectrum of AI inference tasks, which entail executing trained models with efficiency. China’s leading technology companies are prioritizing domestic chip manufacturing in response to the rising demand for AI computing. A report highlighted that Chinese companies are targeting a threefold increase in the nation’s AI chip production by 2026.
The initiative arises in the context of escalating tensions with the US regarding technology exports. Earlier this year, the Trump administration imposed restrictions on the sale of Nvidia’s high-performance H20 chips in China, although the ban was lifted in July. However, Chinese regulators subsequently expressed concerns regarding potential “information leaks” and backdoor access in Nvidia’s chips. Authorities have reportedly advised major firms, including Tencent, ByteDance, and Alibaba, to suspend H20 orders, leading local companies to step in and fill the gap. In response, Chinese companies including Huawei, Semiconductor Manufacturing International Corporation, MetaX, Cambricon Technologies, and Alibaba have intensified their initiatives to enhance chipmaking capacity.
Despite the swift advancements, experts assert that China continues to lag significantly behind US chipmakers. Sources within the industry informed the market that the nation’s manufacturing facilities continue to be unable to produce chips that match the advanced technology of those from the US, which are subject to stringent export restrictions imposed by Washington. According to sources, even Nvidia’s H20, the most powerful chip permitted for sale in China, is considerably less powerful than the H100 or Blackwell series. The lack of advanced manufacturing technology persists in hindering China’s capacity to bridge the divide in high-performance AI hardware.
