We are constantly exploring new AI technologies to enhance your chat experience. Following comprehensive testing and encouraging outcomes, we are excited to introduce an upgraded default model, evolving from llama-1 to the more advanced llama-3.

During our testing phase, we observed a notable 15% reduction in negative feedback and a 7% increase in the regeneration selection rate—both key indicators of improved model performance.

For those who prefer the interaction style of our previous model, we are pleased to continue offering it as an option. You can access the legacy model via the Generation Settings (chat page menu) on our web platform.

Additional Updates for 'I'm All In' Tier Users:

• We have discontinued the MythoMax feature due to its low usage. Although it will still appear in our mobile apps, the upgraded default model will be utilized.
• The default model now supports an expanded 8K token context exclusively for 'I'm All In' tier users.

We are committed to continually improving your experience and invite your feedback on these changes.