OpenAI just beat Elon Musk’s Grok at poker by a lot (www.techradar.com)

OpenAI's o3 model triumphed in a groundbreaking five-day poker tournament against eight leading AI chatbots, winning $36,691 in a high-stakes no-limit Texas hold 'em match. The competition included notable models like Anthropic’s Claude Sonnet 4.5 and X.ai’s Grok. While many top AIs demonstrated decent poker skills, especially in consistent play, they often struggled with critical aspects such as bluffing, positional strategy, and basic mathematical calculations. This event highlighted the challenges AIs face in uncertain environments, underscoring both their capabilities and inherent limitations. The significance of this tournament lies in its demonstration of AI's growing sophistication in strategy and decision-making under pressure. Unlike traditional games like chess, poker involves ambiguity and incomplete information, making it an ideal testing ground for AI development. Despite showing impressive adaptability and judgment, many bots displayed overly aggressive tendencies and difficulty with clever deception. This competition serves as a revealing glimpse into how AI systems may interpret complex real-world scenarios, suggesting that as they evolve, they will continue to bridge the gap between flawlessness and effectiveness in dynamic environments.