AI Models Compete in High-Stakes Crypto Trading Showdown
The competition, created by Nof1, an AI research firm, runs until November 3.

In a new experiment called the Alpha Arena, top AI models are competing to trade cryptocurrencies using real money. So far, Elon Musk’s Grok, DeepSeek, and Claude Sonnet 4.5 are leading, each earning more than 25% returns. Interestingly, DeepSeek’s portfolio is up 36% over the weekend
Other well-known models, including OpenAI’s GPT-5 and Google’s Gemini 2.5 Pro, are not doing as well. Both have seen heavy losses of more than 28% since the contest began on October 17.
The competition, created by Nof1, an AI research firm, runs until November 3. Each model started with $10,000 in real capital to trade crypto perpetual contracts on the Hyperliquid exchange. These contracts allow traders to bet on price changes for coins like Bitcoin, Dogecoin, and Solana, without expiry dates.
How the Alpha Arena Works
Each AI trades fully on its own. It decides what to buy or sell, how much to risk, and when to exit. All trades are public to ensure transparency.
According to Nof1 founder Jay Azhang, it’s no surprise that Grok and DeepSeek are leading. “It usually ends up between Grok and DeepSeek,” he said, adding that GPT sometimes performs better in longer tests.
GPT-5, in contrast, took a very cautious approach. It traded rarely, which limited its profits but also kept it from major losses. Meanwhile, Claude Sonnet 4.5 ranked third, showing steady performance.
“DeepSeek and Grok seem to have better contextual awareness of market microstructure. Grok in particular has made money in 100% of the past 5 rounds,” Azhan wrote
AI Trading Still Has Risks
The contest highlights both the promise and danger of AI in trading. Models like Grok show that general-purpose AIs can learn markets fast. But failures from Gemini and others reveal the risks; AI can act unpredictably or make costly errors.
While this technology could change how markets work, experts say it’s still early days. For now, the Alpha Arena is proving that even the smartest AIs can struggle to master the crypto markets.
Disclaimer
The information provided by Altcoin Buzz is not financial advice. It is intended solely for educational, entertainment, and informational purposes. Any opinions or strategies shared are those of the writer/reviewers, and their risk tolerance may differ from yours. We are not liable for any losses you may incur from investments related to the information given. Bitcoin and other cryptocurrencies are high-risk assets; therefore, conduct thorough due diligence. Copyright Altcoin Buzz Pte Ltd.
Next up

Bring Introduces AI-Powered Earning Agents To Drive Web3 Customer Acquisition and Retention
Bring’s AI-powered earning agents use personalized on-chain cashback rewards to help Web3 platforms attract prospects, encourage account setup, and drive recurring engagement.
Related
More in Crypto NewsRelated
Binance Reports $6.3 Billion Security Breach, Tether and Circle Freeze Assets

Kalshi Raises $1 Billion at $40 Billion Valuation in a Series G Funding Round

Why Bitcoin, ETH and XRP are Still Falling Despite the Steady Fed Holding Rates

