Performance Gap Between Frontier and Open-Weight AI Models Reaches 29 Elo Points

The performance disparity between frontier AI models and open-weight alternatives has expanded to 29 Elo points. This trend highlights the increasing difficulty for open-source models to keep pace with the rapid advancements of closed-source systems.