When you buy through affiliate links in our content, we may earn a commission at no extra cost to you. Learn 
Today, Google announced it is rebranding its Bard chatbot and releasing Gemini Ultra, its most powerful large language model (LLM).
As of today, users can interact with Gemini online or via Google Play and the App Store in over 40 languages. The base version of the chatbot uses the Gemini Pro 1.0 model.
“Our mission with Bard has always been to give you direct access to our AI models, and Gemini represents our most capable family of models. To reflect this, Bard will now simply be known as Gemini,” Google’s announcement blog post said.
At the same time, Google announced the release of Gemini Advanced, an upgraded version of the chatbot available via a Google One Premium Plan ($20 per month), which has access to Ultra 1.0.
The model is designed to perform better at highly complex tasks like coding, logical reasoning and following instructions while supporting longer conversations.
Key Takeaways
- Google rebrands its Bard chatbot as Gemini and launches Gemini Ultra, its most powerful large language model (LLM) yet.
- Gemini supports over 40 languages and integrates tightly with Google’s ecosystem of products.
- Gemini Advanced looks to excel in tasks like coding and reasoning and supports text, voice, and image inputs.
- This move positions Gemini as Google’s flagship AI product, and the competition between Google and OpenAI-Microsoft intensifies.
What Do the Bard Rebrand and Gemini Advanced Launch Mean for the AI Market?
The rebrand of Bard to Gemini marks an attempt by Google to position the Gemini family of LLMs as its flagship, multimodal AI product. The newly released Gemini supports a range of inputs, including text, voice, and images. Thanks to Google’s Imagen 2 text-to-image model, it also can generate images.
One of the core reasons for this is that multimodality is becoming a competitive necessity in the LLM race. After all, we’ve already seen providers like OpenAI and Microsoft incorporating multimodality into their product portfolios.
For instance, OpenAI added an integration with the text-to-image model DALL-E 3 and the ability to support voice inputs at the end of last year, and just yesterday, Microsoft announced it was adding the ability to create and edit images with Copilot, launching a new model called Deucalion.
Introducing Gemini Advanced
However, perhaps the most significant disruptive influence in the market now is Gemini Advanced, which not only offers access to Google’s powerful Ultra 1.0 model but also gives access to 2TB of storage in the cloud.
In addition, the organization says that Premium subscribers can use Gemini via Gmail, Docs, Slides, and Sheets and access data included within those apps.
When considering this, Gemini Advanced’s access to a more powerful multimodal AI model and deep integration with the Google product ecosystem at the same price point as the Plus version of ChatGPT makes it a much more competitive product against solutions from OpenAI and Microsoft.
Of course, Microsoft still has a very competitive offering in its own right — not just with Bing Chat using GPT-4, but with Copilot integrating with Office 365 apps like Word, PowerPoint, Excel, Teams, and Power. Likewise, OpenAI still has time to drop the long-awaited GPT-5.
Is Gemini Advanced Better Than ChatGPT?
Based on the information we have available, Gemini Advanced demonstrates some impressive performance, even outperforming ChatGPT, more specifically the GPT-4 version, in certain areas.
Most notably, according to Google’s research, the Ultra 1.0 model it uses is the first LLM to outperform human experts on the massive multitask language understanding (MMLU) benchmark with a score of 90%. By comparison, GPT-4’s score on the same benchmark was 86.5%.
Google’s research also reports that Ultra achieves state-of-the-art results in 30 out of 32 benchmarks measured, including popular text and reasoning benchmarks, image understanding benchmarks, speech recognition, and speech translation benchmarks.
That being said, while Gemini Ultra performs on a similar level to GPT-4, it doesn’t clearly surpass it.
Ethan Mollick, an Associate Professor at the Wharton School of the University of Pennsylvania who has been experimenting with the model for six weeks, noted in a blog post that “Gemini Advanced is clearly a GPT-4 class model,” but also noted it “does not obviously blow away GPT-4 in the benchmarks.”
In terms of differentiation, Mollick argued that “GPT-4 is much more sophisticated about using code and accomplishes a number of hard verbal tasks better,” whereas “Gemini is better at explanations and does a great job integrating images and search.”
In any case, it’s clear that the gap between OpenAI and Google’s models has closed to a razor’s edge with a viable alternative backed with the right product ecosystem to challenge the monolithic OpenAI-Microsoft partnership, which recently triggered an inquiry from the European Commission.
The Bottom Line
The LLM market just got much more interesting. With Bard giving way to Gemini and Gemini Advanced going head to head with ChatGPT Plus, the competition between Google and OpenAI-Microsoft is getting much higher.
If there’s any company that can unseat OpenAI as the dominant provider in the market, it is Google, just through the sheer power of its experienced DeepMind team and the wealth of proprietary data gathered from products like Google Search and YouTube.
But it is too early to say that GPT-4 has been dethroned just yet.
