Google Unveils Next-Gen Gemini 1.5 AI Model

Google has announced the launch of its latest artificial intelligence model, Gemini 1.5, marking a big leap forward in the company’s AI technology.

The new model is described as making considerable strides in various dimensions, including the ability to process information more efficiently and over longer contexts than its predecessor, Gemini 1.0 Ultra. The demo videos Google shared today illustrate just how big a change Gemini 1.5 is.

“Last week, we rolled out our most capable model, Gemini 1.0 Ultra, and took a significant step forward in making Google products more helpful, starting with Gemini Advanced. Today, developers and Cloud customers can begin building with 1.0 Ultra too — with our Gemini API in AI Studio and in Vertex AI,” said Sundar Pichai, CEO of Google and Alphabet.

The introduction of Gemini 1.5 comes as Google continues to prioritize safety and rapid progress in the development of its AI models. According to Pichai, Gemini 1.5 Pro achieves comparable quality to the 1.0 Ultra model while using less computational power, therefore being more efficient.

Gemini 1.5 Pro can handle a standard context window of 128,000 tokens, but a version limited to select developers and enterprise clients can handle up to 1 million tokens (the smallest units of data an AI model needs to process).

What does this mean? Google says “1.5 Pro can process vast amounts of information in one go — including 1 hour of video, 11 hours of audio, codebases with over 30,000 lines of code or over 700,000 words. In our research, we’ve also successfully tested up to 10 million tokens.” That’s pretty impressive.

Check out the demo below where Gemini 1.5 Pro examines a 402-page PDF transcript from the Apollo 11 mission to the moon. Google even gave a brief drawing of a picture of a boot about to step down and it was able to answer it was the moment Neil Armstrong set foot on the moon. AI is going to run the world, folks.

“Longer context windows show us the promise of what is possible. They will enable entirely new capabilities and help developers build much more useful models and applications,” Pichai added.

Here’s Gemini 1.5 Pro taking on a 44-minute silent Buster Keaton movie (696,000 tokens), able to analyze plot points and other events to mention time codes, and even answer multi-modal prompts (such as giving a brief sketch and asking what the scene is):

Earlier this month, Google launched Gemini in Canada, finally (previously known as Bard).

With the rapid pace of AI models from Google and OpenAI–what will the AI future look like in five years?