As part of the December update announcement for Pixel smartphones, Google detailed a bit of Gemini, its new and improved AI model. In a separate announcement, the company explained Gemini in full, filling us all in on why we should be so hyped for it.
Google describes Gemini is its most flexible AI model yet, capable on running on anything from data centers to your Pixel phone. “Its state-of-the-art capabilities will significantly enhance the way developers and enterprise customers build and scale with AI,” they explain in a blog post.
I’m not savvy in the AI speak, so here’s Google to explain performance. Full performance details can be viewed here.
With a score of 90.0%, Gemini Ultra is the first model to outperform human experts on MMLU (massive multitask language understanding), which uses a combination of 57 subjects such as math, physics, history, law, medicine and ethics for testing both world knowledge and problem-solving abilities.
Our new benchmark approach to MMLU enables Gemini to use its reasoning capabilities to think more carefully before answering difficult questions, leading to significant improvements over just using its first impression.
There’s three models for Gemini: Gemini Ultra, Pro, and Nano. Ultra is the largest and most capable on the bunch, while Pro is designed to scale and Nano is most efficient. For us smartphone nerds, we’re concerned about Gemini Nano, which Google announced for Pixel 8 Pro, powering features like Summarize in the Recorder app and Smart Reply in Gboard (available first to WhatsApp users).
Google notes that Android developers can get their hands on Gemini Nano via AICore, a new system capability available starting in Android 14. You can sign up right here if interested.
And it isn’t stopping here. Google says Gemini will be available in products and services like Google Search, Ads, Chrome, and Duet AI. Buckle up, I guess.