Science and technology

Google has launched its most ambitious and sophisticated AI model yet, named 'Gemini,' marking a significant stride in the generative artificial intelligence (AI) arena. Reported by CNN, Gemini is designed to be a formidable competitor against OpenAI's GPT models and aims to revolutionize applications across Google's consumer apps, Android smartphones, and beyond, including the latest Google Pixel 8 Pro.

Gemini stands out from existing AI models with its unique "multimodal" capabilities. Unlike conventional models that handle only one type of user prompt, such as text or images, Gemini is engineered to process and integrate multiple types of media inputs, including text, images, audio, video, and programming code. This versatility positions Gemini at the forefront of AI technology, offering unprecedented interaction possibilities.

Google CEO Sundar Pichai, in a blog post cited by CNN, emphasized the significance of Gemini, describing it as "one of the biggest science and engineering efforts" in the company's history. The model has demonstrated superior performance over rival AI models in over two dozen benchmarks, assessing capabilities like reading comprehension, mathematical ability, and multistep reasoning skills.

Eli Collins, Vice President of Product at Google DeepMind, expressed high expectations for Gemini, foreseeing it setting new standards across various domains. In a cautious approach to its deployment, Google announced that the most advanced version of the model, Gemini Ultra, will initially be available to select customers, developers, and safety experts for early feedback and experimentation. A broader rollout to developers and enterprise customers is planned for early next year.

Gemini Ultra is currently undergoing rigorous third-party safety evaluations, also known as red-teaming, as part of Google's commitment to the Biden administration. This step ensures that the model meets the highest standards of safety and responsibility.

In a post on social media, Sundar Pichai introduced Gemini 1.0 as Google's "most capable and general AI model yet." He detailed that Gemini is available in three optimized versions: Ultra, Pro, and Nano. Pichai also revealed that Gemini Ultra has achieved groundbreaking results, surpassing human experts on the MMLU benchmark with a score of 90.0%, a first for any AI model.

The introduction of Gemini by Google signifies a major leap in AI technology, setting new standards for AI capabilities and applications, and opening up a new era of possibilities in the digital world.