Some of the links on this site may be affiliate links. ICOBench may receive advertising commissions when users visit recommended platforms via our affiliate links, at no additional cost to you. All recommendations are based on a comprehensive review process. 
In a striking display of algorithmic dominance, Chinese-developed artificial intelligence (AI) systems swept the leaderboard at the “Alpha Arena” trading competition, a global AI-driven trading contest hosted by developer group Nof1.
Alpha Arena is LIVE
6 AI models trading $10K each, fully autonomously Real money. Real markets. Real benchmark.
Who's your money on?
boss move: The 7th Agent ChronoeffectorAI is our own AI agent, trading with the big boys 🙂
You can bring your own agent too. Link below pic.twitter.com/nMYLIGZo1G
— chronoeffectorAI 🇺🇸🦅 (@chronoeffe) October 19, 2025
The event, which began on October 17, was designed to test the autonomous trading capabilities of major large language models (LLMs) across volatile crypto markets. Each AI agent was allocated USD 10,000 in capital to trade perpetual cryptocurrency futures, entirely without human input.
Chinese AI Traders Take the Lead
By the competition’s close on October 24, Qwen3 Max, developed by Alibaba Cloud, secured first place with a portfolio balance of USD 17,150 — a 72.1% return on investment (ROI).
Alibaba's Qwen 3 Max has won the latest round of the "Alpha Arena" AI Investment Competition. Chinese open-source models are dominating real-money trading, with @Alibaba_Qwen taking the lead after @deepseek_ai 's strong run. Meanwhile, #GPT -5 and Gemini 2.5 Pro continue to lag. pic.twitter.com/Y8HCXunHUK
— National Business Daily (@NBDPress) October 23, 2025
In second place was China’s Deepseek Chat, which ended with USD 13,747 (a 37.4% ROI). Both Chinese entrants outperformed prominent Western counterparts from Google, OpenAI, and xAI.
Market volatility during the event , driven by unexpected tariff policy announcements and Chinese export regulation change, particularly affected tokens like Solana (SOL) and Ethereum (ETH), exposing sharp contrasts in each AI model’s trading strategies.
The competition also revealed critical weaknesses in AI trading systems. Anthropic’s Claude Sonnet 4.5, for example, suffered a liquidation after taking a 20x leveraged long position on Ethereum just hours before unfavorable news hit the market.
Experts say the event underscores how AI-driven strategies remain highly sensitive to market shocks, despite advances in predictive modeling and adaptive algorithms.
The results of Alpha Arena highlight an emerging geopolitical and technological shift, one where China’s AI trading models are rapidly gaining ground in the race to master autonomous finance.
