However, the public launch of Gemini has been delayed due to issues with handling non-English queries. Despite this, Google maintains that Gemini has strong multilingual capabilities. The AI model also boasts multimodal performance, meaning it can understand different types of information, such as text, images, video, and audio. A paid version of the chatbot running on Gemini Ultra, called Bard Advanced, is expected to be released early next year.
Key takeaways:
- Google has unveiled its new AI model, Gemini, which is the company's most powerful AI to date and will be integrated into various Google services.
- Gemini comes in three tiers - Gemini Ultra, Gemini Pro, and Gemini Nano, with the Ultra version touted as a competitor to OpenAI's GPT-4.
- Google claims that Gemini outperforms GPT-4, scoring over 90% on the industry standard benchmark MMLU, and excels in multimodal performance, comprehending different kinds of information such as text, images, video, and audio.
- Google's chatbot Bard will be running on a finely tuned version of Gemini Pro, with a paid version running on Gemini Ultra expected to roll out early next year.