Google Set to Release Gemini Ultra Next Week

Google will release Gemini Ultra by next week, according to a document posted by a user on X. Further, Google is set to rename Bard as "Gemini," mirroring the name of its underlying model. 

The changelog provided by Android app developer Dylan Roussel is dated February 7, and it explicitly mentions that the paid Gemini Advanced tier will be released simultaneously. Additionally, there is a reference to the upcoming release of a Gemini app for Android.

To access Gemini Ultra, users will need to subscribe to Gemini Advanced, a paid plan available in over 150 countries and territories. Gemini Ultra will include multimodal capabilities, enhanced coding features, and the ability to analyze documents, files, and data.

In the changelog screenshot shared by Dylan Roussel, there is mention of the upcoming release of the Gemini App for Android. Integrated with Google apps such as Gmail, Maps, and YouTube, the Gemini App streamlines tasks on your phone. It offers interaction through text, voice, or images for added convenience 

To chat with Gemini on Android users can download the Gemini app in the Google Play Store. On iOS, try Gemini in the Google app.

Google released Gemini last year in December. According to the technical paper Google published, Ultra outperforms GPT-4 in 7 out of 8 benchmark tests and is the first model to surpass human experts in MMLU (massive multitask language understanding). However, it should be noted that it uses COT@32 instead of 5-shot learning.

Ultra also outperforms GPT-4V (vision) in all 10 benchmarks for image interpretation. Google recently upgraded Bard to Gemini Pro. This enhanced version of Bard enables text-to-image generation, powered by the latest Imagen-2 model, which is claimed to outperform OpenAI’s DALL·E2.