Google made waves in the AI world by unveiling its new multimodal model, Gemini.
This advanced AI system aims to outperform the current leader, OpenAI’s GPT-4, across a variety of tasks involving text, images, audio, video and more.
So what makes Gemini so special?
For starters, it was built from the ground up to seamlessly operate across multiple data types. That’s a key advantage over other models like GPT-3 and GPT-4 which started as text-only systems and later added on image and audio capabilities.
Gemini breezed through a battery of benchmark tests, outperforming GPT-4 in most areas. This includes complex reasoning, reading comprehension, basic math, Python coding, image recognition and even translating speech.
The top-of-the-line Gemini Ultra variant even bested GPT-4 in an IQ-style test spanning 57 academic subjects. It scored 90% accuracy compared to GPT-4’s 86.4%.
Now, don’t get too excited just yet.
The powerful Gemini Ultra won’t be available to the public until sometime next year. But a slimmed-down version called Gemini Pro has already been integrated into Google’s conversational AI assistant, Bard.
This gives Bard a major boost in capabilities, perhaps finally making it a legitimate rival to OpenAI’s wildly popular ChatGPT.
You can try out the upgraded Bard right now.
The launch of Gemini signals Google’s serious commitment to leading the AI race. With continued advances, Gemini may soon become the most versatile and capable AI system ever created.
But only time will tell whether it can dethrone OpenAI as the dominant player in this fast-evolving landscape.
