From Future plc: March 26, 2024

Large language models (LLMs) like OpenAI’s GPT-3.5 are making waves in the world of AI, with GPT-3.5 emerging as the champion of Street Fighter III battles at the SF3 LLM Colosseum. At the Mistral AI Hackathon, LLMs were found to outperform reinforcement learning algorithms in predicting player moves. (Word count: 50)

The LLMs calculate player moves in Street Fighter III battles based on screen descriptions, previous moves, opponent actions, and health bars. OpenAI’s GPT-3.5 proved its prowess in intense battles, surpassing Mistral LLM versions and preview models like GPT-4. Source code for the project is available on Github for gaming enthusiasts. (Word count: 50)

The potential applications of LLMs in gaming are vast, raising questions about the future of AI-powered gaming experiences. While the focus is on gaming now, some speculate about the military applications of large language models like GPT-3.5, known for its powerful capabilities in war game scenarios. (Word count: 50)

Read more at Future plc: OpenAI’s GPT-3.5 is the champion of the Street Fighter III LLM Colosseum, beating Mistral on its home turf