Google, a subsidiary of Alphabet Inc., has rolled out an array of upgraded artificial intelligence (AI) features for its cloud-computing clients. This move is part of Google’s efforts to catch up with competitors, particularly Microsoft Corp. and startup OpenAI, who have capitalized on the growing demand for AI solutions.
The latest offering, Gemini Pro, is designed for enterprises, enabling developers to create applications using Google’s state-of-the-art AI model, Gemini. This large-scale AI system, introduced last week, leverages vast datasets to generate new content based on user requests. Google emphasizes that Gemini Pro will be initially free for cloud customers, with plans to ensure competitive pricing in the future.
Gemini Pro’s text-based capabilities are touted to be four times less expensive for input and two times less expensive for output compared to the previous AI model, PaLM 2, released in June. According to Google Cloud CEO Thomas Kurian, Gemini is built to seamlessly understand and operate across different types of information, including text, code, audio, image, and video – mirroring the way humans process diverse information simultaneously.
While Google has been a pioneer in generative AI, its products have faced challenges in popularity. To address this, Google introduced Gemini, offering three versions: Ultra, Pro, and Nano. The Nano version runs directly on devices like the Pixel 8 Pro smartphone. Gemini Pro, specially designed for Bard – Google’s AI chatbot, competes with OpenAI’s ChatGPT.
Google aims to dispel the notion of trailing behind OpenAI by releasing Gemini Pro to app developers and businesses. The goal is to position itself as a formidable player in the evolving AI landscape, challenging OpenAI’s latest system, GPT-4, available through Microsoft’s Azure cloud-computing platform.
Gemini Pro supports 38 languages across 180 countries, currently accepting text as input and generating text as output. Additionally, Google introduces Gemini Pro Vision, a platform capable of handling text- and image-based prompts from users. The largest and most capable model, Gemini Ultra, will be available for select cloud customers and partners for early experimentation before a public release next year.
Google plans to integrate Gemini Pro into key cloud products, including Google AI Studio and Vertex AI. Google AI Studio, a web-based developer tool, facilitates fast app development with the Gemini API. Vertex AI offers more customization options, allowing businesses to use their data to build Gemini-powered tools such as search engines and chatbots.
To make Gemini Pro more accessible, Google has significantly improved its pricing. Developers can access Gemini Pro and Gemini Pro Vision for free through Google AI Studio, while Vertex AI remains free until early next year.
Additionally, Google announced upgrades to existing AI models: Imagen 2, focusing on text-to-image technology with enhanced photorealism, text rendering, and logo generation capabilities. MedLM, a family of models fine-tuned for the healthcare industry, builds on the success of Med-PaLM 2, an AI model trained on specialized medical knowledge.
In a global partnership with Mistral AI, a Paris-based AI startup, Google will distribute some AI products, including optimized proprietary language models, on its cloud infrastructure. Although non-exclusive, the collaboration reflects a shared commitment to open-source software and responsible AI principles. Mistral AI’s CEO, Arthur Mensch, praised Google Cloud’s flexibility, infrastructure capabilities, and commitment to privacy and security, aligning with Mistral AI’s mission to develop openly available models.
