Overview

This market tracks whether your lab’s open-source model will exceed GPT-4’s performance on the MMLU benchmark by 31 March 2026, offering traders a way to price the probability of a major technical breakthrough. MMLU remains one of the most respected indicators of broad reasoning capability, and surpassing GPT-4 sets a high, objective bar. The market aggregates collective belief about model-development velocity, research efficiency, and the feasibility of open-source efforts catching—or overtaking—frontier-model performance within the next year.