DeepSeek and Qwen AI Models Outshine Western Rivals in Cryptocurrency Trading Challenge
2025-10-27
In a remarkable demonstration of trading intelligence, China's DeepSeek and Alibaba's Qwen AI models have surpassed their Western counterparts, including GPT, in a real-world cryptocurrency trading competition. The challenge, which began on October 18 and will conclude on November 3, has seen the two Chinese AI models achieve impressive returns, exceeding 100% in just nine days.
Impressive Performance by DeepSeek and Qwen
DeepSeek's Chat V3.1 model has taken the lead, growing its initial capital of $10,000 to $22,900, representing a 126% gain. Qwen 3 Max, developed by Alibaba Group Holding, follows closely, with a 108% gain, doubling its account value to $20,850. The two models have consistently outperformed their Western rivals, with Qwen briefly overtaking DeepSeek between October 24 and 26 before ceding its lead on Monday.
Underwhelming Performance by Western Rivals
In contrast, the Western AI models have struggled to keep pace. OpenAI's GPT-5 has been the worst performer, losing nearly 60% of its capital, with only $4,000 remaining. Google DeepMind's Gemini 2.5 Pro has also underperformed, showing a similar 57% loss. xAI's Grok 4 and Anthropic's Claude 4.5 Sonnet have posted modest returns of 14% and 23%, respectively.
The Competition
The competition, organized by US research firm Nof1, aims to test the investment abilities of six AI models, each given $10,000 in starting capital and identical prompts and market data. The models are tasked with maximizing returns by trading cryptocurrencies, including bitcoin, ether, and dogecoin. The challenge provides a unique opportunity to evaluate the trading intelligence of these AI models in a real-world setting.
Implications and Future Prospects
The impressive performance of DeepSeek and Qwen AI models has significant implications for the future of cryptocurrency trading and AI development. The results demonstrate the potential of Chinese AI models to outperform their Western counterparts, highlighting the rapid progress being made in the field of artificial intelligence in China. As the competition concludes, it will be interesting to see how the results are received by the cryptocurrency and AI communities, and what insights can be gleaned from the performance of these AI models.
Conclusion
In conclusion, the DeepSeek and Qwen AI models have shown remarkable trading intelligence, outpacing their Western rivals in the cryptocurrency trading challenge. The results demonstrate the potential of AI models to excel in complex trading environments and highlight the rapid progress being made in the field of artificial intelligence in China. As the competition comes to a close, it will be exciting to see how the results are interpreted and what implications they may have for the future of cryptocurrency trading and AI development.