Google recently unveiled its new artificial intelligence (AI) system called Gemini, positioning it as a rival to OpenAI’s popular ChatGPT chatbot. This Gemini vs. ChatGPT article compares Gemini and ChatGPT on history, availability, pricing, data sources, and more to determine which AI assistant looks more promising.
What is Gemini AI?
Google Gemini is the latest generation of large language models (LLMs) developed by Google DeepMind. It is the successor to LaMDA and PaLM 2, and it is designed to be more powerful and capable than its predecessors in several ways.
Also Read: ChatGPT Mobile apps surpass 110M installs, nearly $30M in revenue from the apps
Google has introduced three iterative versions of its Gemini large language model (LLM), collectively marketed as “Gemini 1.0”—the inaugural generation of the company’s neuronal network-driven AI. These include:
- Gemini Ultra: The most capable variant (albeit slower runtimes)
- Gemini Pro: An adaptable, widely deployable model now powering Bard
- Gemini Nano: The most efficient iteration, sacrificing some functionality for on-device integration
For now, Gemini Pro serves as the backbone for Google’s fledgling Bard conversational service. Gemini Nano is embedded into the Google Pixel 8 smartphone, enabling handy features like productive summaries and smart replies.
Gemini Ultra, the most advanced version, is still in development. Google plans to launch it in 2024 to power a more advanced iteration of Bard.
Beyond Bard, Gemini will also extend into Search, Ads, Chrome, and other Google products over the next two years.
Gemini vs. ChatGPT
Google boldly claims its upcoming Gemini Ultra AI can surpass ChatGPT across most performance metrics. However, the currently available mid-tier Gemini Pro underpinning Bard so far fails to support this lofty promise.
Also Read: Gemini: Google launches new AI model to compete with ChatGPT
Early analysis shows Gemini Ultra does narrowly best ChatGPT-4 in certain head-to-head tests. But this advanced iteration won’t debut until 2024. In the interim, reviewers are pitting the Pro variant against ChatGPT’s free GPT-3.5 model, and initial reactions are lukewarm at best.
For example, The Verge’s Nilay Patel posted Gemini hallucinated a fictional interview when asked an important question. Similarly, Matteo Wong from The Atlantic, acknowledges that Gemini can outperform GPT-4 in select categories but calls improvements merely “iterative” rather than revolutionary.
Tech writer Ryan Morrison also conducted a head-to-head evaluation of ChatGPT versus Gemini published on Tom’s Guide. His comparative analysis pitted ChatGPT’s broadly available GPT-3.5 model against Google’s fledgling Gemini Pro, currently underpinning Bard.
Granted, even esteemed ChatGPT suffers occasional AI lapses into hallucination and requires ongoing refinements to fix exploits. Yet the onus remains on Google to truly showcase Gemini’s brilliance in winning over ChatGPT users—a feat presently unmatched despite the company’s aggressive promotional claims.
Pricing
Much of Gemini’s usage will be baked into free Google products like Search, YouTube, and Chrome. So users can access some basic Gemini features at no cost.
That said, Google will likely charge for more advanced Gemini integrations in certain enterprise services down the line. The pricing models aren’t yet clear.
Also Read: ChatGPT has new voice and image recognition superpowers
In contrast, ChatGPT offers a clear and simple pricing structure. There’s a free tier for basic usage. For $20 per month, ChatGPT Plus provides additional features, higher usage limits, and faster response times. And ChatGPT Enterprise allows businesses to customize the AI as needed.
For now, Gemini’s pricing remains nebulous compared to ChatGPT’s transparent options. But cost likely won’t be a core factor in choosing one AI over the other for most everyday users.
Models and Data Sources
The current Bard platform uses both Google’s previous Bard LLM AI model and the new Gemini Pro model. Over time, more Google products will incorporate next-gen Gemini models alongside existing AI systems.
Gemini Ultra, set to launch next year, will likely become Google’s flagship AI offering. Early test results show it surpasses ChatGPT-4 across a range of benchmarks, including text, image, and video understanding.
“Our most-capable model, Gemini Ultra, advances the state of the art in 30 of 32 benchmarks, including 10 of 12 popular text and reasoning benchmarks, 9 of 9 image understanding benchmarks, 6 of 6 video understanding benchmarks, and 5 of 5 speech recognition and speech translation benchmarks.” –Jeff Dean, Chief Scientist, Google DeepMind
As for data sources, both ChatGPT 3.5 and GPT-4 models are trained on publicly available data from the internet, with content cutoffs around September 2021. In contrast, details remain scarce on what datasets were used to develop the Gemini models.
Availability
One can already start testing Google’s new Gemini AI through the recently launched Bard chatbot website. Unlike some beta launches, you don’t have to wait in line or join a waitlist to try Gemini.
The middle-level Gemini Pro version currently powers Bard’s English-language chatbot in over 170 countries globally. Similarly, people worldwide have had open access to try ChatGPT since it was first launched to the public in November 2022.
However, both companies limit some advanced features only to paying users. Google seems to be saving its future Gemini Ultra and Nano upgrades for premium subscriptions or products down the road.
And ChatGPT charges $20 per month for ChatGPT Plus to remove usage caps and get faster response times compared to the free tier. So in short – basic Gemini and ChatGPT features are now easy to access, but you’ll likely need to pay extra later on if you want to unlock the full power of the AI as this technology battle unfolds.
