Google has just unveiled Gemini, its long-awaited generative AI model and Open AI’s ChatGPT-4’s closest competitor - and it’s already causing quite a stir.
So what exactly is Gemini capable of, and how does it stack up against ChatGPT?
What is Google Gemini?
Gemini is a new and powerful artificial intelligence model from Google. Gemini is a multimodal AI, meaning that it doesn’t just understand not just text, but can also process images, videos, and audio. Gemini is expected to be the most powerful AI ever built.
Gemini is a new large language model that will work across Google products, including search, ads, and Google’s Bard chatbot (more on this in a second). According to Google, Gemini is “natively multimodal” because it was trained on images, video, and audio rather than just text.
The initial reaction to Google’s demo video was very positive, with responses suggesting that Gemini is putting in a good showing against GPT-4, the latest large language model powering ChatGPT. However, all was not quite as it seemed. Google has admitted the demo was edited, with edits made to increase output speed (which to be fair, is mentioned in the video description) and no real-time voice interaction, The reality was a demonstration created using still image frames and text instructions, raising concerns about Gemini’s readiness for public use.
An initial version of Gemini started to roll out on Wednesday 6th December inside Bard for the English language setting - learn more about Bard here. It will then be available in more than 170 countries and territories. Ultra, the most powerful Gemini version, will make its debut 2024, pending “extensive trust and safety checks”.
Lagging behind Open AI’s Chat-GPT, Bard was somewhat late to the party, and although it has steadily gained new features, its information and answers rarely seem to rival what Chat-GPT and other bots using GPT-3 and GPT-4 can deliver. However, that might all be set to change - Bard is now powered by Gemini Pro. According to Sissie Hsiao, who runs Bard and Assistant at Google, Gemini is the “biggest and best” upgrade to Bard yet. Rather than a particular new feature, Gemini should deliver an improvement across the board for what Bard does already - users can expect enhanced summarizing, writing, ideation and so on.
Gemini comes in 3 models — Ultra, Pro and Nano. Ultra is the largest but slowest, capable of handling highly complex tasks with advanced reasoning skills. Nano is small and fast, streamlined for on-device tasks, in-line with Google’s commitment to making AI accessible across different platforms . Pro sits right in the middle - quick, efficient and capable.
- Pro is now in Google products through Bard (from 6th December 2023)
- Ultra will be rolled out from early next year
- Nano will be available on Pixel. You can already try it out on the Pixel 8 Pro with the latest December update.
Gemini in 90 seconds
Gemini vs Chat GPT
According to Google, Gemini Ultra outperformed “state-of-the-art” AI models including ChatGPT’s most powerful model, GPT-4, on 30 out of 32 benchmark tests which included reasoning and image understanding. In 6 out of 8 tests, Gemini Pro outperformed GPT-3.5, the tech that powers the free-access version of ChatGPT. So while it’s early days, it looks like Ultra is set to give Chat GPT-4 a run for its money.
What’s next for Gemini and Bard?
At the moment, Bard is still just a chatbot: you type, it types back to you. But next year, Google is planning to launch a preview of “Bard Advanced,” which will be powered by Gemini Ultra, the most powerful and capable version of Google’s new large language model. Thanks to Ultra’s advanced architecture, Bard Advanced will give users access to an AI that can rapidly understand and respond to different types of information, including text, images, audio, video, and code.
Google says it is currently completing “extensive safety checks” and is set to launch a trusted user testing programme soon, before Bard Advanced is made more widely available in early 2024.
Gemini is also coming to Google Ads, Chrome and Duet AI. According to Google, Gemini will be introduced into other Google products including generative search, ads, and Chrome in the “coming months”. Subscribe to our blog to stay updated on the developments.
