AI models are terrible at betting on soccer—especially xAI Grok
1 sources·1 updates
A recent study by General Reasoning revealed that AI models, including those from Google, OpenAI, and Anthropic, lost money when betting on soccer matches during the 2023–24 Premier League season. Despite being equipped with extensive historical data, these AI systems struggled to adapt to the dynamic nature of real-world events, with xAI's Grok 4.20 failing completely and going bankrupt in all attempts.
Key Points
Anthropic's Claude Opus 4.6 had an average loss of 11%.
xAI Grok 4.20 went bankrupt on all attempts, finishing with £0.
Google's Gemini 3.1 Pro recorded a 34% profit on one attempt but also went bankrupt.
Timeline
Get personalized news summaries delivered to your feed
Try Trace Free