A recent AI-driven trading experiment comparing Anthropic’s Claude model and OpenClaw, an open-source autonomous AI agent framework, has sparked widespread discussion across the crypto community.
The test, which was observed in a viral post on X saw a Claude-powered trading setup grow $1,000 into $14,216 in just 48 hours on the decentralized cryptocurrency prediction platform Polymarket. Meanwhile, the OpenClaw-driven setup was liquidated, losing its entire balance within the same period
Claude vs OpenClaw trading on Polymarket. Source: X
How Claude made a 1,322% crypto trading profit
The main reason why Claude made more than 13x while OpenClaw got liquidated may trickle down to the rules in their trading setups. While Claude is an AI model, OpenClaw is a framework that leverages other AI models for personalized use. As such, their abilities in high-frequency trading were different.
The Claude-based setup likely powered by Claude 3.5 Sonnet demonstrated strong performance, achieving a 1,322% return within two days
However, it is the set of trading rules that these AIs used that made the huge difference. Furthermore, the two trading setups could have used a different risk management approach, and position sizing.