Cheryl Wu, an Econ PhD student at Yale and research scholar at GovAI, stated that Chinese frontier models remain 4-5 months behind US models on the Effective Compute Index and that the gap shows no signs of closing. She noted Chinese models now meet most coding needs, which may explain perceptions of faster progress. In a separate post Wu compared pricing and found Chinese models cheaper below an ECI score of 155 while US models cost less above that level. Researchers including Jaime Sevilla retweeted the ECI observations.