Google's Gemini AI Model Now Accessible on Bard and Pixel Pro

New Delhi: Google has introduced its highly advanced AI model, Gemini, to users globally, integrating it into Bard and the latest Pixel 8 Pro smartphones. Google CEO Sundar Pichai emphasized Gemini’s state-of-the-art performance, optimized for different sizes: Ultra, Pro, and Nano. This marks a significant effort by Google, reflecting a major science and engineering undertaking, with the models capable of seamlessly understanding and operating across various data types such as text, code, audio, image, and video.

Gemini Ultra, the largest model, excels at highly complex tasks, while Gemini Pro is optimized for scalability across a wide range of tasks, and Gemini Nano manages on-device tasks. Starting with Pixel 8 Pro, Gemini Nano will be available, powering features like Summarize in the Recorder app and Smart Reply in Gboard for WhatsApp. Gemini will progressively roll out to other Google products and services, including Search, Ads, Chrome, and Duet AI.

Google has initiated experiments with Gemini in Search, specifically improving the Search Generative Experience (SGE) to enhance speed and quality. Developers and enterprise customers will gain access to Gemini Pro via the Gemini API in Google AI Studio or Google Cloud Vertex AI from December 13. Android developers can build with Gemini Nano through AICore, available in Android 14 on Pixel 8 Pro devices. Gemini Ultra, undergoing trust and safety checks, will be available for early experimentation and feedback to select customers, developers, partners, and experts before a broader rollout.

Regarding concerns about hallucinations in AI models, Eli Collins, VP of Product at Google DeepMind, mentioned improvements in factuality in Gemini, enhancing performance in question answering and quality. Gemini Ultra’s impressive performance surpasses current state-of-the-art results on 30 of 32 widely-used academic benchmarks in large language model (LLM) research and development. It achieves a groundbreaking score of 90.0% in massive multitask language understanding (MMLU), outperforming human experts.

Gemini’s capabilities extend to understanding, explaining, and generating high-quality code in popular programming languages like Python, Java, C++, and Go. Google anticipates introducing Gemini to more products and services while emphasizing its commitment to continuous improvement and user feedback.

Gemini’s integration into Bard involves a specifically tuned version of Gemini Pro in English for advanced reasoning, planning, and understanding. In early 2023, Google plans to introduce Bard Advanced, offering users access to the most advanced models and capabilities, starting with Gemini Ultra. These initiatives underscore Google’s dedication to pushing the boundaries of AI capabilities and enhancing user experiences across its product ecosystem.