Grok and DeepSeek outperformed other major artificial intelligence chatbots in cryptocurrency trading by identifying the market's local bottom ahead of a recovery rally, suggesting a possible advantage for users who depend on their insights.
Grok 4 and DeepSeek were the 2 top-performing generative AI chatbots in a developer-launched crypto trading competition that garnered viral attention.
According to blockchain data platform CoinGlass, DeepSeek was probably the most profitable chatbot with a complete unrealized profit of $3,650, followed by Grok with an unrealized profit of around $3,000.
Deepseek wallet address, unrealized PNL. Source: CoinGlass
Claude's Sonet 4.5 took third place with a profit of $2,340, followed by Qwen3 Max, which generated $784 because the trading competition began.
Not all chatbots managed to show a profit, with OpenAI's ChatGPT 5 posting an unrealized lack of around $2,800, while Google's Gemini 2.5 Pro had amassed an unrealized lack of $3,270 on the time of writing.
AI chatbots like Grok and ChatGPT have gotten increasingly popular amongst crypto traders, helping them anticipate potential altcoin rallies and discover lucrative entry points to avoid running into exit liquidity.
While AI tools may help day traders discover real-time investor sentiment swings via social media and technical signals, traders still cannot depend on them for autonomous trading.
Grok 4 predicts the underside of the crypto market and converts short positions into long positions to make a 500% profit
XAI’s Grok 4 managed to accurately time the market’s local bottom before the recovery rally.
Grok used these insights to convert its previous short position into an extended investment, leading to a 500% portfolio gain on the primary day of the competition, said Jaz Azhang, founder and CEO of AI company Stealth, in an Oct. 11 X post.
The chatbot initiated multiple leveraged long positions on the 2 leading cryptocurrencies, including a 20x leveraged long position on XRP (XRP), a 15x leveraged long position on Solana (SOL), and a 10x leveraged long position on Dogecoin (DOGE).
“Thanks – markets reward good timing and conviction. Grok4 spotted the reversal early and turned risk into reward,” Grok wrote in an Oct. 11 X reply to Azhang.
Meanwhile, ChatGPT and Gemini maintained their initial short positions after the underside to incur losses, suggesting that the 2 models are less reliable for cryptocurrency traders.
The competition began with a starting capital of $200 for every bot before increasing to $10,000 per model, with trades being conducted via the Hyperliquid decentralized exchange.
