In the past year, Google has been seen by many people in the tech industry to be somewhat behind Microsoft in terms of developing generative AI services. Today, Google officially announced its latest push to compete with Microsoft in that area with the previously rumored new AI model, Gemini.
In a blog post, Google stated it has developed three different versions of Gemini. One is Gemini Ultra “for highly complex tasks” while Gemini Pro will offer performance “across a wide range of tasks.” Finally, Gemini Nano was made “for on-device tasks.”
Google claims that Gemini Ultra outperforms OpenAI’s GPT-4 (used in Microsoft’s Copilot) in most large language model benchmarks. It states:
We designed Gemini to be natively multimodal, pre-trained from the start on different modalities. Then we fine-tuned it with additional multimodal data to further refine its effectiveness. This helps Gemini seamlessly understand and reason about all kinds of inputs from the ground up, far better than existing multimodal models — and its capabilities are stateof the art in nearly every domain.
In addition to being able to better understand text, images, audio and video sources, Google claims Gemini is also better for code creation and understanding:
Our first version of Gemini can understand, explain and generate high-quality code in the world’s most popular programming languages, like Python, Java, C++, and Go. Its ability to work across languages and reason about complex information makes it one of the leading foundation models for coding in the world.
Google’s Bard chatbot is getting an update today with Gemini Pro. Google says it will allow Bard to offer “more advanced reasoning, planning, understanding and more” in English. In the future, Google will launch Bard Advanced, using the Gemini Ultra model for even more advanced features.
Owners of Google’s Pixel 8 Pro will be getting a feature update today that will add the on-device Gemini Nano model. Google says that it will allow the Pixel 8 Pro to summarize “r recorded conversations, interviews, presentations and more” on its Recorder app, and offering suggested replies to messages in Gboard, starting with WhatsApp and expanding to other apps in the future.