Google has announced the availability of its AI model Gemini Pro for developers and enterprises to craft solutions for their specific needs. This move is accompanied by a commitment to ongoing refinement based on user feedback.
Gemini comes in three distinctive sizes: Ultra, Pro, and Nano. The Nano version has already been integrated into Android – starting with the Pixel 8 Pro – while a specially tuned iteration of Gemini Pro is embedded in Bard.
Empowering developers
Key features of Gemini Pro include its outperformance in research benchmarks compared to similarly-sized models, a 32K context window for text (with plans for larger windows in future versions), and competitive pricing. The model supports 38 languages across 180+ countries and territories.
Gemini Pro’s capabilities span a range of functions, including embeddings, semantic retrieval, custom knowledge grounding, and chat functionality. The model accepts text as input and generates text as output, with a dedicated Gemini Pro Vision multimodal endpoint available for processing both text and imagery.
Developers can now leverage SDKs for Gemini Pro, supporting Python, Kotlin, Node.js, Swift, and JavaScript.
Google AI Studio – a free web-based developer tool – provides a seamless experience for prompt development and obtaining API keys for app development. This tool offers a generous free quota of 60 requests per minute, significantly exceeding other free offerings.
Transitioning to Vertex AI on Google Cloud
For those seeking a fully-managed AI platform, the transition from Google AI Studio to Vertex AI is smooth. Vertex AI enables customisation of Gemini with full data control, incorporating additional Google Cloud features for enterprise security, safety, privacy, and data governance.
Developers using Vertex AI can fine-tune and distil Gemini with their company’s data, build search and conversational agents, and deploy knowing that Google Cloud does not use customer data for model training.
As of now, developers enjoy free access to Gemini Pro and Gemini Pro Vision through Google AI Studio. Vertex AI developers can also test these models at no cost until general availability early next year, after which charges will apply per 1,000 characters or per image.
Google plans to launch Gemini Ultra – its most advanced model for highly complex tasks – in early 2024.
(Image Credit: Google)
See also: David DeSanto, GitLab: AI’s impact on software development in 2024
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is co-located with Digital Transformation Week.
Explore other upcoming enterprise technology events and webinars powered by TechForge here.
