Google brings Gemini AI to developers and organizations worldwide
The version comes with a 32 KB context window for text, and future versions will have a larger context window.
San Francisco: Google announced Wednesday that it is making its large language model (LLM) Gemini Pro available to developers and enterprises to help them build their own use cases for advanced tasks. The first version of Gemini Pro is now accessible via the Gemini API. The version comes with a 32 KB context window for text, and future versions will have a larger context window. It’s now free to use, within limits, and it will be competitively priced. It supports 38 languages in more than 180 countries and territories worldwide.
The company also introduced Google AI Studio, a free, web-based developer tool that allows them to quickly develop prompts and then get an API key to use in developing their app.
“You can sign in to Google AI Studio with your Google Account and take advantage of the free quota, which allows 60 requests per minute – 20x more than other free offerings,” the company said in a statement.
Currently, developers can access Gemini Pro and Gemini Pro Vision for free through Google AI Studio, with up to 60 requests per minute, making it suitable for most app development needs.
Developers of Vertex AI (leading enterprise AI platform) can try the same models, with the same pricing caps, for free until general availability early next year, at which point you will be charged per 1,000 characters or per image in Google AI Studio and Vertex AI , the company said.
“Starting today, Gemini is part of a vertically integrated and vertically optimized AI technology stack made up of several key components – all designed to work together,” said Thomas Kurian, CEO of Google Cloud.
Google Cloud provides industry-leading AI-optimized infrastructure for businesses, the same used by Google, to train and operate models.
“We offer you this infrastructure in our cloud regions as a service, to run in your Google Distributed Cloud data centers and at the edge. Our entire AI infrastructure stack is built with system-level co-design to increase efficiency and productivity when training, tuning and operating AI,” said Kurian.
Duet AI is Google’s AI-powered collaboration partner that helps users use Google Workspace and Google Cloud.
For example, Duet AI in Google Workspace helps users write, create images, analyze spreadsheets, compose and summarize emails and instant messages, and summarize meetings.
“Early next year we will launch Gemini Ultra, our largest and most capable model for highly complex tasks, after further refinement, safety testing and gathering valuable feedback from partners. We will also be bringing Gemini to more of our developer platforms like Chrome and Firebase,” Google said.