Core Insights - The experiment involved six advanced AI models trading in the cryptocurrency market with a starting capital of $10,000 each, aiming to determine which model could generate the highest returns [1][2] - The Chinese model Qwen 3 Max emerged as the top performer, achieving a net value of $15,594, representing a nearly 60% increase, while other models like GPT-5 and Gemini 2.5 Pro suffered significant losses of 72.1% and 61.9% respectively [1][10][11] Group 1: AI Performance - Qwen 3 Max and DeepSeek were the standout performers, with Qwen achieving a net value of $15,594 and DeepSeek at $12,926 by October 24 [10][11] - In contrast, GPT-5 and Gemini 2.5 Pro faced severe declines, with net values dropping to $2,783 and $3,792 respectively, indicating a loss of over 60% of their initial capital [11][12] Group 2: Trading Strategies and Personalities - The AI models exhibited distinct trading styles, with DeepSeek characterized as aggressive, focusing on high-frequency trading and short-term gains [12][13][16] - Qwen 3 Max adopted a more conservative approach, executing only 22 trades with a focus on low volatility and risk management, which contributed to its success [17][18] - In contrast, Gemini 2.5 Pro engaged in excessive trading with high leverage, leading to chaotic decision-making and significant losses [22][24] Group 3: Market Dynamics and AI Adaptability - The experiment highlighted that intelligence alone does not guarantee profitability in trading; adaptability and the ability to self-correct in response to market conditions are crucial [25][26] - The financial market serves as a dynamic environment that tests the judgment, adaptability, and risk management capabilities of AI models, contrasting with more static environments like games [26][27]

Core Insights - Alibaba's Qwen 3 Max won the first place in the AI investment competition "Alpha Arena," outperforming other models including GPT-5, which ranked last [1] - The competition, initiated by the US research lab nof1.ai, serves as a public test for observing AI's autonomous trading capabilities [1] Group 1: Competition Overview - "Alpha Arena" features six major AI models including DeepSeek, GPT-5, Gemini 2.5 Pro, Claude Sonnet 4.5, Grok 4, and Alibaba's Qwen 3 Max [1] - Each model started with an initial capital of $10,000 and engaged in perpetual contract trading on the decentralized exchange Hyperliquid [1] - The competition's sole evaluation criterion is the return on investment, with all trades being publicly visible [1] Group 2: Performance Highlights - Alibaba's Qwen 3 Max currently leads the profitability rankings, demonstrating a stable upward trend in its return curve [1] - The investment strategy of Qwen 3 Max shows its ability to continuously self-optimize in high-frequency market feedback through real-time reinforcement learning [1] - DeepSeek, another Chinese open-source model, maintained a significant lead for a considerable duration before the competition [1]

Core Insights - Nof1 conducted a three-day trading competition among six top AI models, each given $10,000 to trade in the decentralized exchange Hyperliquid's perpetual contracts market [4][5][6] - The participating models included Claude 4.5 Sonnet, DeepSeek V3.1 Chat, Gemini 2.5 Pro, GPT 5, Grok 4, and Qwen 3 Max [4] Performance Summary - DeepSeek's success was attributed to a clear and strictly executed trading strategy, effectively diversifying investments across six major cryptocurrencies like Ethereum (ETH) and Bitcoin (BTC) to mitigate risks from price volatility [9][10] - DeepSeek employed moderate leverage and set clear stop-loss points for each trade, allowing for quick exits from losing positions while letting profitable trades run [10] - Grok 4 followed closely with a 30% return, while other models underperformed due to various execution errors, such as order failures and missed trading signals [10] - Some models misinterpreted strategies, with overly cautious approaches missing market opportunities, while others took aggressive positions in rising markets, leading to rapid capital drawdowns [10] - Alpha Arena highlighted that the differences in model performance stemmed from their execution of instructions, risk management, and trading capabilities [10]

A new real-market cryptocurrency trading experiment that pits leading artificial intelligence models against one another to evaluate their respective investing abilities has seen a DeepSeek model outperform rivals so far. In Alpha Arena, launched on Friday by US research firm Nof1, six large language models (LLMs) were given US$10,000 each to invest in six cryptocurrency perpetual contracts on the decentralised exchange Hyperliquid, including bitcoin and solana. As of 2pm on Tuesday, DeepSeek's V3.1 had ...

Core Insights - Jay Chou's recent troubles involve a Bitcoin account managed by his magician friend, Cai Weize, who claimed the account was locked a year ago, resulting in a loss of funds [1][2] - The article discusses the emergence of AI models competing in the cryptocurrency market, highlighting a competition called Alpha Arena where six top AI models are trading cryptocurrencies [3][4] Group 1: AI Competition Overview - The competition involves six AI models, each given $10,000 to trade perpetual contracts on the Hyperliquid platform, with trading pairs including BTC, ETH, BNB, SOL, XRP, and DOGE [4][6] - The performance of these AI models is measured by risk-adjusted returns, focusing not only on profits but also on the risks taken [6][7] Group 2: Performance of AI Models - As of the latest update, DeepSeek Chat V3.1 leads with an account value of $14,310 and a return of +43.1%, showcasing a strategy of high leverage and concentrated positions [11][12] - Grok 4 follows with an account value of $13,921 and a return of +39.21%, employing a high-leverage long-only strategy [12][21] - Claude Sonnet 4.5 has an account value of $12,528 and a return of +25.28%, focusing on a conservative trading approach [12][23] - In contrast, GPT-5 and Gemini 2.5 Pro are underperforming, with returns of -24.78% and -27.74% respectively, indicating poor trading strategies and high transaction costs [12][30] Group 3: AI's Role in Investment - The article emphasizes that AI's greatest value in investment may lie in transparency, allowing investors to see trading records and decision-making processes, unlike human-managed accounts [40][41] - The ambition behind the AI competition is to use financial markets as a training ground for AI, aiming for continuous learning and adaptation to market dynamics [34][35]

Core Insights - A competition was held by nof1.ai where six leading language models were given $10,000 each to trade in a real market environment, aiming to maximize risk-adjusted returns [1][6] - The models involved include Claude 4.5 Sonnet, DeepSeek V3.1 Chat, Gemini 2.5 Pro, GPT 5, Grok 4, and Qwen 3 Max, all trading perpetual contracts on Hyperliquid [1] Performance Summary - After approximately 60 hours of trading, DeepSeek emerged as the best performer with a total portfolio value of nearly $14,000, achieving a return of about 40% [3] - Grok 4 followed closely with a portfolio value around $13,300, both models profiting primarily from long positions in Bitcoin and Ethereum [5] - Claude and Qwen focused on trading Ripple and Ethereum, ranking third and fourth respectively, while both outperformed Bitcoin's spot market [6] - GPT 5 and Gemini showed significant losses, with portfolio values of $7,300 and $6,900, indicating losses of approximately $2,700 and $3,100 [6] Market Context - The competition aims to provide a benchmark that closely resembles real-world conditions, with financial markets being ideal due to their dynamic, adversarial, open, and highly unpredictable nature [6] - Analysts suggest that there is significant potential for killer applications in the DeFi + AI space, indicating a strong interest in the involvement of LLMs in on-chain gaming [7]