New Model:

Based on positive responses to the addition of Magnum-12B, we're making the big brother of this model Magnum-72B available for I'm All In subscribers.

We had not updated our Advanced Models since the addition of WizardLM2-8x22 at the beginning of the summer. This new model is based on the large-size model Qwen2-72B model (Alibaba) which outperformed other leading open-source models in 15 benchmarks, including language understanding, language generation, multilingual capability, coding, mathematics, and reasoning. Multilingual capabilities include 27 languages, in addition to Chinese and English.

Removed Model: We have finally retired our legacy Spicy2 model that was based on Llama2. The low usage wasn't making it worthwhile to bear the cost of keeping that model running. We have also removed the availability of Command-R which is getting replaced by Magnum-72B. Other Updates:

• Generation of sexually explicit images of celebrities and public figures is no longer possible using our Avatar Generator.
• We have fixed an issue where newer openai api keys were not properly saved in the advanced settings section of the profile. The longer keys are now properly saved on web. The fix to mobile applications shold be available next week.