The initial version, Gemini 1.5 Pro, is already out for testing. Sundar Pichai, Google’s CEO, shared in a blog post that 1.5 Pro matches the quality of 1.0 Ultra but uses less computing power.

Sundar Pichai Unveils Advanced Gemini Ultra: Gemini AI

Pichai highlighted the smarter features of Gemini 1.5, emphasizing its ability to comprehend a vast amount of information simultaneously. The model can now handle up to 1 million pieces of information at once, a remarkable improvement.

Also Read: Sundar Pichai Unveils Advanced Gemini Ultra

Google’s CEO mentioned that this allows 1.5 Pro to process extensive data, including one hour of video, eleven hours of audio, codebases with over 30,000 lines, or more than 700,000 words. In their research, Google successfully tested up to 10 million tokens.

What can Gemini 1.5 Do?

Gemini 1.5 employs a new Mixture-of-Experts (MoE) architecture, praised by Demis Hassabis, CEO of Google DeepMind, for its efficiency. The MoE architecture uses only the necessary parts of the model’s brains for the task at hand, making it work better and faster.

Hassabis explained in a blog post that these advancements in model architecture enable Gemini 1.5 to learn complex tasks more quickly and maintain quality while being more efficient to train and serve. Google’s teams are benefitting from these efficiencies, allowing them to iterate, train, and deliver more advanced versions of Gemini faster than before.

Also Read: Sundar Pichai Unveils Advanced Gemini Ultra

During testing, Google presented the Gemini 1.5 Pro model with a 44-minute silent movie, and it reportedly accurately analyzed plot points, events, and even small details that might be overlooked.

How can you use Gemini 1.5?

Currently, Gemini 1.5 is accessible only to developers and enterprise users. Google has not disclosed when and how it will be made available to the public. However, in its announcement blog, Google assured its commitment to responsibly bringing each new Gemini model generation to billions of people, developers, and enterprises worldwide.

Google mentioned that when releasing Gemini 1.5 Pro to a broader audience, it will have a standard context window of 128,000 tokens. The company plans to offer different pricing levels, starting with the standard window size and going up to 1 million tokens as they improve the model.

During the testing phase, early users can try out the 1 million token window for free, but Google cautions that response times may be longer with this new feature.

Also Read: Sundar Pichai Unveils Advanced Gemini Ultra

Get the Business News along with India’s Latest Updates at TheNewsay and Follow for Live Updates – Facebook, Instagram, Threads, Twitter, Pinterest, KooApp and YouTube.