The Gemini family of artificial intelligence (AI) models has received a number of improvements from Google, one of which is the new 1.5 Flash model, which is a lightweight AI model designed for efficiency and speed.
On Google AI Studio and Vertex AI, the Gemini 1.5 Pro and 1.5 Flash AI models are both accessible in public preview with a context window containing one million tokens.
Gemini models are used by more than 1.5 million developers now across all of our products. The CEO of Google and Alphabet, Sundar Pichai, said, “You’re using it to debug code, get new insights, and build the next generation of AI applications.”
We are still in the early stages of the platform transition in AI. We see a plethora of opportunities for developers, entrepreneurs, innovators, and everyone else in the future,” he said.
Additionally, the business announced that Gemini 1.5 Pro is being integrated into Workspace applications and other Google products, such as Gemini Advanced.
The business claimed, “We’re also sharing a glimpse of ‘Project Astra, a look at the future of universal AI agents,’ and announcing our next generation of open models, ‘Gemma 2’.”
By continually encoding video frames, fusing audio and video input into an event chronology, and storing this data for effective recall, these AI agents are made to digest information more quickly.
According to Google, “These agents can respond quickly in conversation and better understand the context in which they’re being used.”
In the meanwhile, the Gemini 1.5 Flash is more affordable to service and is better suited for large-scale, high-volume, high-frequency applications.
Gemini 1.5 Flash, according to Hassabis, is excellent in summarizing, chat programs, captioning images and videos, extracting data from large documents and tables, and more.
Additionally, Google unveiled Imagen 3, its best text-to-image model to date, and Veo, its newest and most sophisticated video production model.
“Veo is our most advanced approach for creating videos to date. Eli Collins, Vice President, Product Management, said, “It produces excellent 1080p resolution videos in a variety of cinematic and visual styles that can last longer than a minute.”
The model offers an unmatched degree of creative control as it is also capable of understanding terminology used in film, such as “timelapse” and “aerial shots of a landscape.”
In addition, the business introduced a new conversational experience, a context window with one million tokens, tools that “let Gemini take action on your behalf,” and Gemini 1.5 Pro to Gemini Advanced members in more than 35 languages.
The business unveiled a new Gemini model that is specifically designed for Google Search in an effort to improve it even more.
“It combines our best-in-class Search systems with Gemini’s advanced capabilities, including multi-step reasoning, planning, and multimodality,” said the tech behemoth.
According to Liz Reid, VP and Head of Google Search, “we’re making ‘AI Overviews’ available to everyone in the U.S., with more countries coming soon.” This change goes into effect today.
Google’s Search Labs initiative, “AI Overviews,” has already been utilized billions of times by users.