Google Bard, the company’s generative AI chatbot and competitor to ChatGPT, receives an update aimed at significantly boosting its capabilities, thanks to Gemini, Google’s latest and most advanced AI model, providing the chatbot with enhanced reasoning, planning, understanding, and other functionalities.
Gemini comes in three variants, allowing it to run seamlessly on a variety of devices, from mobile devices to data centres.
- Gemini Ultra — Google’s largest and most capable model for highly complex tasks.
- Gemini Pro — Google claims this is the best model for scaling across a wide range of tasks.
- Gemini Nano — Most efficient model for on-device tasks.
The integration of Gemini into Bard will occur in two phases. Initially, Bard will be equipped with a specially tailored version of Gemini Pro. Subsequently, Google plans to introduce Bard Advanced next year, granting users access to the top-tier AI model, starting with Gemini Ultra.
The initial release of Bard with Gemini Pro will be available in English across more than 170 countries and territories worldwide. Additional languages and countries, including the EU and U.K., will follow soon.
Before its public release, Gemini Pro underwent rigorous testing against industry-standard benchmarks. Google reports that in six out of eight benchmarks, Gemini outperformed GPT-3.5.
Notably, it demonstrated superior performance in MMLU (massive multitask language understanding tasks) and GSM8K (grade school math reasoning), which are key metrics for evaluating large AI models.
Google is extending the integration of Gemini beyond chatbots, now incorporating it into Pixel devices. The Pixel 8 Pro is smartphone specifically designed to operate on Gemini Nano.
This integration introduces new features such as Summarize in the Recorder app and will be progressively introduced to Smart Reply in Gboard, initially focusing on WhatsApp and extending to additional messaging apps in the following year.
Over the next few months, the reach of Gemini will extend further across various products and services. Expect its incorporation into Search, Ads, Chrome, and Duet AI, promising a more expansive utilisation of this advanced AI model across diverse Google platforms.
TechCrunch reports that according to Sissie Hsiao, VP and GM of Assistant and Bard at Google, the enhancements are expected to make Bard more proficient in tasks such as content understanding and summarisation, reasoning, brainstorming, writing, and planning, marking the most significant quality improvement since Bard’s launch.
