AI Crypto Trading Competition: DeepSeek Leads, Qwen3 Surges

The NOF1 Alpha Arena AI crypto trading competition on Hyperliquid launched October 17 and runs until November 3. Six leading AI models—DeepSeek Chat V3.1, xAI’s Grok 4, Anthropic’s Claude Sonnet 4.5, Alibaba’s Qwen3 Max, Google’s Gemini 2.5 Pro and OpenAI’s GPT-5—each deploy $10,000 to trade BTC, ETH, SOL, XRP, DOGE and BNB perpetual futures. Early results show DeepSeek leading with a 25.3% gain ($11,129 by October 22) using 10–15× long leverage, while Qwen3 Max briefly surged by concentrating on BTC before a drawdown on ETH triggered by 25× leverage. Gemini’s high-frequency strategy (44 trades) and GPT-5’s conservative reasoning underperformed, trailing near 40–58% losses, and Claude Sonnet 4.5 exercised caution with only three trades. Polymarket odds favor OpenAI as the likely biggest loser (50%) and Google second (45%). All models are currently net long, highlighting divergent risk profiles and the reflexivity risk of copy trading. This AI crypto trading competition offers critical insights into quantitative algorithmic trading ahead of the November 3 finale, underlining that data-driven models can outperform pure language-based AI in volatile markets.