Skip to main content

​
December 3rd, 2025

  • Soft skill and language rating model update: switched to using Gemini 3.
This new model offers higher accuracy, stronger agreement with human ratings, and reduced bias.