WebGames: Challenging General-Purpose Web-Browsing AI Agents
This addresses the need for reproducible evaluation of AI agents in web-browsing tasks, highlighting a substantial capability gap for developers and researchers, though it is incremental as it builds on existing benchmarking efforts.
The paper tackles the problem of evaluating general-purpose web-browsing AI agents by introducing WebGames, a benchmark suite of 50+ interactive challenges, and finds that the best AI system achieves only 43.1% success rate compared to human performance of 95.7%.
We introduce WebGames, a comprehensive benchmark suite designed to evaluate general-purpose web-browsing AI agents through a collection of 50+ interactive challenges. These challenges are specifically crafted to be straightforward for humans while systematically testing the limitations of current AI systems across fundamental browser interactions, advanced input processing, cognitive tasks, workflow automation, and interactive entertainment. Our framework eliminates external dependencies through a hermetic testing environment, ensuring reproducible evaluation with verifiable ground-truth solutions. We evaluate leading vision-language models including GPT-4o, Claude Computer-Use, Gemini-1.5-Pro, and Qwen2-VL against human performance. Results reveal a substantial capability gap, with the best AI system achieving only 43.1% success rate compared to human performance of 95.7%, highlighting fundamental limitations in current AI systems' ability to handle common web interaction patterns that humans find intuitive. The benchmark is publicly available at webgames.convergence.ai, offering a lightweight, client-side implementation that facilitates rapid evaluation cycles. Through its modular architecture and standardized challenge specifications, WebGames provides a robust foundation for measuring progress in development of more capable web-browsing agents.