What Is Gemini Advanced & How Does It Compare To Chat-GPT?

Gemini Advanced was previously known as Google Bard and is the premium version of the AI model. The premium version was released on Friday and is now available to the public to purchase. You can access the current and free version of Gemini here - https://gemini.google.com/app.

Google apparently held off releasing the premium version because of security concerns but they obviously feel as though it’s in a great position and probably felt as though they were losing market share to Chat-GPT, which would over time be harder and harder to win back, even when it came to die-hard Google fans. 

As part of the Gemini Advanced launch, Google has launched its most capable AI model, Ultra 1.0. When Bard originally launched just under a year ago it used a large language model called LaMDA. In May 2023, this was upgraded to utilise its PaLM2 language model and in December, Google upgraded Bard to use its Gemini Pro AI model. The free version of Gemini is still using the Gemini Pro AI model under the hood but if you pay for Gemini Advanced, you’ll get access to Ultra 1.0.

You can upgrade your Google One subscription to include access to Gemini Advanced. You can check what Google One plan you have here - https://one.google.com/plans. By upgrading to their most premium plan you’ll have access to Gemini Advanced (along with 2TB of cloud storage, VPN and several other benefits) for the next 2 months free of charge, rising up to £18.99 / month thereafter. 

After upgrading, when you visit gemini.google.com to access the AI assistant, you can switch between "Gemini" and "Gemini Advanced" in a drop-down menu in the upper-left corner of the web interface (similar to switching between GPT-3.5 and GPT-4 in Chat-GPT)

Gemini will also be available in a new Gemini app for Android and in the Google app on iOS. launched in the upcoming weeks.

What is Google One? 

Google One began as a cloud storage service intended for the consumer market but is now roping in AI capabilities as part of its membership perks. It gives you the ability to back up data on your mobile phone, store photos safely in the cloud, provides your devices with a VPN for safer browsing, 

Is Gemini Advanced Included In My Google Workspace Subscription? 

Gemini Advanced is only purchasable through Google One so your subscription to Google Workspace won’t give you access to Gemini Advanced. 

Can I Use It In Google Workspace? 

At time of writing you can’t directly use it from within Google Workspace apps such as Docs, Gmail, Sheets, Slides, etc, however there are plans to release it so it will be accessible in Workspace. A Workplace extension is available within Gemini Advanced extensions and that is able to access your workplace apps. 

What Can It Do?

It can do essentially anything that GPT-4 can do. Gemini Advanced is a standalone app but over the coming weeks will be coming to Google Workspace apps like Gmail, Docs, Slides, and Sheets. Within Google Workspaces and Google Cloud it was known as “Duet AI” and will now become "Gemini for Google Workspace and Google Cloud." Just like the previous Duet AI, Gemini will be able to assist you with composing emails in Gmail, analysing data in Sheets, or summarising content in Docs.

Much like with other language models you can perform highly complex tasks such as coding, logical reasoning, following nuanced instructions and collaborating on creative projects. You can type a question, talk or add an image for all kinds of help. You can take a picture of your flat tire and ask for instructions, generate a custom image for a dinner party invitation, or ask for help writing a difficult email. 

In a recruitment context it can help your team to structure and write your job advertisements & role descriptions, providing them with a specific tone, promoting inclusivity and optimising them with relevant keywords so they rank better in search engine results. It can help to remove gender bias from emails, ads, role specifications, contracts, etc. It can help consultants when matching CVs with live roles and ranking candidates based on relevancy. Gemini can help to draft candidate outreach emails and LinkedIn messages. It could help create general and client specific interview guides, using the client’s website as a reference point while also generating a list of questions they may encounter in the interview process. You could pull the candidate salary data from your CRM into a CSV file or sheets and ask Gemini to generate a salary survey for you. And lastly a big one in the recruitment sector currently would be the ability to use it to power a basic chatbot on your recruitment website.

How Do They Compare?

At a glance, it looks like Google is finally starting to catch up with OpenAI's GPT-4 Turbo model when it comes to capability. I’ve been using the free version of Bard over the last couple of months but noticed it wasn’t even close to GPT-4’s capability.  

Like ChatGPT-4, Gemini Advanced is multimodal, which means you can upload images and discuss them with the chatbot. It can visit links on the web, and it can also generate images using Google's Imagen 2 model. It cannot generate images yet, however I’m sure these plans are in the pipeline. Google Gemini seems to be able to access more websites with its browsing feature than Chat-GPT. This is likely down to the fact that so many websites have blocked OpenAI’s crawlers, including several major media owners and news organisations. Gemini remains largely free to index the web, likely due to Google's position as the most popular search engine.

Both systems have safety guardrails as we’ve probably all seen but each of the two are triggered in different ways. Gemini seems more willing to do darker writing when compared with GPT-4 but it’s been noticed that Gemini tends to throw out a few more refusals than ChatGPT-4. Just this morning I asked it to check the spelling and grammar in my blog post and it replied, “I can't create images yet so I'm not able to help you with that.” - obviously an irrelevant and unexpected response. 

From what I’ve seen so far GPT-4 is much more sophisticated about using code and was near perfect at one point, the model did drop off for a few months last year but seems to be back on track now. GPT-4 accomplishes a number of hard verbal tasks better. Gemini is better at explanations and does a great job integrating images and search. Although if I were looking to generate AI images I would personally use Midjourney. 

It feels like Gemini Advanced is waiting for some fine-tuning and additional features to be turned on after a period of general use by premium members, but when it does get it right it's easily on par with using Chat-GPT Plus. 

The marketplace for app integrations is much bigger in Chat-GPT when compared to Gemini’s extensions. In Chat-GPT general users are able to create their own “GPTs”, however Gemini Advanced currently only includes Google Flights, Google Hotels, Google Maps, Google Workspace and YouTube. 

The data show that Gemini is currently on par or just slightly below par with GPT-4, but what remains to be seen is just how much better it gets as Google improves the bot in response to user feedback and how it keeps up or eclipse GPT-4.