Google Gemini Pro is Coming to Businesses and Developers

Google is giving businesses and developers the first glimpse of its most powerful large language model, Gemini, through its API. The model is available in three sizes: Ultra, Pro and Nano.

Starting today, the Gemini Pro API is available to developers via Google’s free web-based developer tool, AI Studio (formerly Makersuite). Gemini Pro is also available to enterprises through Google Cloud’s Vertex AI platform. Companies can use it to build applications starting today.

Google said it plans to further fine-tune the model in the coming weeks based on user feedback. “We can’t wait to see what developers and enterprises build with Gemini,” the company said in a blog post.

Gemini Pro is already powering Bard, Google’s answer to ChatGPT. The initial version has a low 32,000 context window for text, which means it can handle around 5,333 words (32,000 tokens). By comparison, GPT-4 Turbo, OpenAI’s newest model, can handle 128,000 tokens. However, Google said later versions of Gemini Pro will have greatly expanded lengths.

Other Gemini Pro features include support for 38 languages, function calling, embeddings, semantic retrieval and custom knowledge grounding.

Currently, it only accepts text as input and generates text as output. However, there is a dedicated Gemini Pro Vision multimodal endpoint that accepts both text and imagery - images and video as input while generating text as output. That is available from today.

Related:Google Launches Gemini Next-Generation Large Language Model

Gemini Pro’s API is currently free to use – but has a maximum of 60 queries per minute. There is a pay-as-you-go version coming soon which is less restrictive, however, with Google saying it will be “competitively priced” as it looks to take on OpenAI.

Google has already released the prices for Gemini Pro: $0.00025 per thousand characters or $0.0025 per image. Output costs $0.0005 per thousand characters.

Inputs and outputs for the free version of the Google Pro API will be used by Google to improve its products, the company admitted, but the paid-for version will not.

Credit: Google

Imagen 2 and other news

Alongside Gemini Pro, Google has other models to add to Vertex, including Imagen 2, the company’s latest AI image generation model. Using the most powerful text-to-image diffusion model built by Google DeepMind to date, Imagen 2 can generate high-quality images and can even be used to create realistic logos for businesses. The model can also render text in multiple languages.

Also added to Vertex AI was MedLM, a family of foundation models fine-tuned for the health care industry. Built upon the Med-PaLM 2 foundation model, MedLM is designed to power health care use cases including medical notetaking and medical question and answering. Currently, MedLM is only available to U.S.-based Vertex users, with plans to expand it to Model Garden in the coming weeks. Google also plans to add Gemini-based models to the MedLM suite “soon.”

Related:12 Language Models You Need to Know

Finally, Duet AI for Developers is now generally available. Designed to help developers build applications, Duet AI is a collaboration tool that can be embedded across Google Cloud interfaces to help with code generation and chat assistance. Gemini is coming to Duet AI over the next few weeks.

Duet AI is also being expanded to security operations, with the collab tool making its way to defenders in a unified SecOps platform.