Crypto
AI Trading Experiment Shows Mixed Results: Only Two Models Profit While ChatGPT Loses 63%
AI lab Nof1 tested six AI models in crypto day trading with $10,000 each. Only DeepSeek (+5%) and Qwen3-Max (+22%) profited, while ChatGPT lost 63%. Models used varied strategies on Hyperliquid but faced high fees and inconsistent results. Conclusion: AI remains a useful assistant, not a reliable autonomous crypto trader.
Anyone venturing into day trading cryptocurrencies like Bitcoin or Ethereum needs to monitor and assess news, price movements, and market sentiment. It seems logical, then, to use artificial intelligence (AI) that never tires and makes unemotional decisions.
At least, that’s how the Artificial Intelligence lab Nof1 envisioned the situation and began an experiment in mid-October. Six popular AI models were each given $10,000 in seed capital. ChatGPT 5, Gemini 2.5 Pro, Claude Sonnet 4.5, Grok 4, DeepSeek v3.1, and Qwen3-Max then had two weeks to increase their capital – but only the latter two actually succeeded.
ChatGPT falls far behind in the crypto trading competition
Nof1 founder Jay Azhang presented the detailed results via X : According to him, ChatGPT suffered a 63 percent loss in crypto trading and had only $3,700 left in its account at the end of the competition.
Grok ended up with a 45 percent loss, Gemini lost 56 percent of its starting capital, and Claude Sonnet lost 30 percent. On the bright side were DeepSeek with a gain of almost 5 percent and $489, and Qwen3-Max with a 22 percent gain, equivalent to $2,200.
AI models design various strategies for trading crypto
All six AI models were allowed to try their luck on the decentralized cryptocurrency exchange Hyperliquid, trading Bitcoin, Ethereum, Solana, Binance Coin (BNB), Dogecoin (DOGE), and/or XRP (Ripple). They could also speculate on price movements. The AIs pursued quite different strategies. Gemini, for example, executed 238 trades, while Claude Sonnet only managed 38.
Claude Sonnet typically held open positions in only one or two of the six available cryptocurrencies, while others like Grok and ChatGPT pushed the limits to their maximum. One finding: Individual AI models like Gemini and Qwen3-Max apparently underestimated Hyperliquid’s fees, resulting in additional costs well over $1,000 in the final result.
Conclusion: Artificial intelligence is still, at best, an assistant for crypto traders
Nof1 plans to organize another AI trading competition soon and use the results of the first round to improve its prompts. The outcome of the first experiment, however, is unlikely to encourage anyone to let ChatGPT, Gemini, Grok, and similar platforms dictate their trading strategies.
AI can already be helpful, though, in chart analysis for Bitcoin and altcoins like Ethereum or Solana. Artificial intelligence has its place as an assistant or sparring partner, but it still seems clearly overwhelmed when it comes to being a crypto trader taking sole responsibility.
__
(Featured image by Levart_Photographer via Unsplash)
DISCLAIMER: This article was written by a third party contributor and does not reflect the opinion of Born2Invest, its management, staff or its associates. Please review our disclaimer for more information.
This article may include forward-looking statements. These forward-looking statements generally are identified by the words “believe,” “project,” “estimate,” “become,” “plan,” “will,” and similar expressions. These forward-looking statements involve known and unknown risks as well as uncertainties, including those discussed in the following cautionary statements and elsewhere in this article and on this site. Although the Company may believe that its expectations are based on reasonable assumptions, the actual results that the Company may achieve may differ materially from any forward-looking statements, which reflect the opinions of the management of the Company only as of the date hereof. Additionally, please make sure to read these important disclosures.
First published in BLCOK-BUILDERS.DE. A third-party contributor translated and adapted the article from the original. In case of discrepancy, the original will prevail.
Although we made reasonable efforts to provide accurate translations, some parts may be incorrect. Born2Invest assumes no responsibility for errors, omissions or ambiguities in the translations provided on this website. Any person or entity relying on translated content does so at their own risk. Born2Invest is not responsible for losses caused by such reliance on the accuracy or reliability of translated information. If you wish to report an error or inaccuracy in the translation, we encourage you to contact us.
-
Crowdfunding2 weeks agoCoinbase Acquires Echo for $375 Million to Expand On-Chain Fundraising Capabilities
-
Crypto3 days agoTether’s $10B Profit Spurs Western Union’s Stablecoin Move, Pressuring Ripple
-
Crypto1 week agoBitcoin Surges Above $115,000 as “Uptober” Momentum Builds Ahead of Key Economic Decisions
-
Crowdfunding2 hours agoCafé Alma Saved: Community Support Revives Jacky Fabiené’s Struggling Business



