Why Are Large Language Models So Terrible at Video Games?
- Published
- Source
- IEEESpectrumAI
Recent advancements in large language models (LLMs) have led to significant improvements in various domains, particularly in coding. However, a notable limitation remains: LLMs struggle to play video games effectively. Despite some successes, such as Gemini 2.5 Pro defeating Pokémon Blue in May 2025, these models often perform poorly compared to human players, making frequent mistakes and requiring specialized software to assist them.
Julian Togelius, director of New York University’s Game Innovation Lab and co-founder of AI game-testing firm Modl.ai, discussed these challenges in a recent interview with IEEE Spectrum. He highlighted that while coding resembles a well-structured game with clear tasks and immediate feedback, video games present a more complex landscape that LLMs have yet to navigate successfully. Unlike games like chess or Go, which have been mastered by AI through retraining, video games vary significantly in mechanics and input requirements, complicating the development of a general game AI.
Togelius pointed out that the lack of comprehensive benchmarks for video games further hinders LLMs' performance. While benchmarks have driven improvements in coding, the diverse nature of video games makes it difficult to establish similar metrics. He noted that current LLMs perform poorly even compared to basic algorithms in gaming contexts, primarily due to insufficient training data and challenges in spatial reasoning.
Despite their coding capabilities, LLMs cannot engage in the iterative process of game development, which involves testing and refining gameplay. This disparity raises questions about the future of AI in mastering video games and its implications for broader AI applications.
Original report
Why Are Large Language Models So Terrible at Video Games?
IEEESpectrumAI · Matthew S. SmithThis briefing is an independently written summary based on publicly available reporting and is provided for industry information and news discovery. The original report and source publication are credited and linked where applicable. RobotToday does not claim ownership of third-party source material.
Rights concerns? Contact [email protected] with the relevant URL and details. We will review the matter and take appropriate action where warranted.
More related news
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
Related Suppliers
BioRobotics Laboratory, Korea University of Technology and Education
Korea, Republic of (South)
South Korean university research laboratory focusing on haptics, exoskeletons, rehabilitation robotics, and autonomous ground vehicles.
McMaster Manufacturing Research Institute
Canada
Hamilton, Ontario applied manufacturing research center at McMaster University; runs a 21,000 sq ft facility with seven research labs.
Robotics Education & Competition Foundation
United States
The REC Foundation is a US nonprofit that runs VEX-based robotics and drone competitions reaching 1.1 million students across 70 countries.
Ortelio Ltd
United Kingdom
Ortelio Ltd is a UK cloud robotics research company founded in 2011 that develops AI platforms giving robots intelligence for service and healthcare tasks.
Parata Systems
United States
Durham, North Carolina pharmacy automation maker acquired by BD in 2022; builds robotic vial-filling and packaging systems.
Pensa Systems
United States
Austin-based AI retail shelf platform using autonomous drones and CV to detect stockouts; customers include J&J, AB InBev, and Circle K.
Human in Motion Robotics
Canada
Vancouver, Canada medical-robotics company founded 2016 builds the XoMotion self-balancing exoskeleton for people with spinal cord injury or stroke.
LG Display
Korea, Republic of (South)
Seoul display maker (KRX:034220, NYSE:LPL), formed 1999, that launched plastic OLED humanoid-robot face panels for industrial settings in 2026.
Roboblock System Co., Ltd.
Korea, Republic of (South)
South Korean robotics distribution and development company founded in 1997; supplies AI robots, drones, humanoid kits, and VR/AR systems internationally.
Robotics needs a service framework.
RSF defines a common language for robot service capability, lifecycle operations, certification pathways, and service-provider networks.
Join the RobotToday community on LinkedIn
Daily robotics news, in-depth analysis, conference highlights, and discussions with professionals worldwide.
