Google Gemini AI is now available for enterprise development

Google today announced that its most powerful and capable generative AI model, Gemini, is now available to enterprises for their app development needs.

Announced last week, Gemini comes in three sizes: Ultra, Pro and Nano. With today’s move, the Sundar Pichai-led company is making the Pro version of the model accessible via API. It can be used for free for now, but there are certain usage limitations, the company wrote in a blog post.

In addition to this, it also made a bunch of other announcements in the AI space, including an upgraded Imagen 2 text-to-image diffusion tool and a family of foundation models fine-tuned for the healthcare industry.

Gemini Pro for developers: What to expect?

The first version of Gemini Pro is available via the Gemini API in the Google AI Studio – which gives developers a web-based developer platform to develop prompts and then get an API key to use in app development. It comes with a 32K context window for text generation, which the company says will be expanded in the future. 

“We’ve also made a dedicated Gemini Pro Vision multimodal endpoint available today that accepts text and imagery as input, with text output,” Google wrote.

In an X post announcing the availability, Pichai pointed out that the Gemini API gives developers access to a full range of features, including function calling, embeddings, semantic retrieval, custom knowledge grounding and chat functionality. It also supports 38 languages across 180+ countries. 

Beyond the AI Studio, Gemini Pro is also coming on Vertex AI, Google Cloud’s end-to-end AI platform that includes tooling, fully-managed infrastructure and built-in privacy and safety features for AI development. This gives developers an option to transition to a fully managed environment whenever needed.

Ultimately, the company plans to learn from developer feedback to fine-tune Gemini Pro and move towards the launch of the bigger Gemini Ultra next year. It has been built for more complex tasks.

Free but with a catch

As of now, Google says, Gemini Pro and Gemini Pro Vision can be accessed for free with a rate limit of up to 60 requests per minute. The same applies to developers using the models on Vertex AI – but only until general availability next year. Google says that the free quota is 20 times more than other offerings and should be suitable for most development needs. 

That said, once the offering is generally available, the company plans to charge per 1,000 characters or per image across both Google AI Studio and Vertex AI.

Specifically, the input price of Gemini Pro is kept at $0.00025 per 1K characters and $0.0025 per image, while the output price for both remains the same at $0.0005 per 1K characters.

As some have observed on X, this is far more than comparable pricing from rivals such as OpenAI’s GPT, since Google is charging “per character,” i.e., each letter or number generated by the AI model, versus OpenAI’s and most other AI companies’ “per token” pricing, wherein a numeric token can be used to represent entire words.