Your everyday work tasks could soon feature a fair few more AI tools thanks to a major expansion from Google Gemini.
The company has unveiled its vision of an enterprise world powered by its Gemini AI model, which it says can benefit companies of all sizes, and also be a huge boost to developers everywhere.
Google recently unveiled Gemini in three “sizes” - Ultra, Pro and Nano, each designed for different use cases and products. Nano has already arrived on the company’s Pixel 8 Pro Android smartphone, and a “specifically tuned” version of Pro is being used for Google Bard. The company is now expanding Gemini Pro to developers, allowing enterprises to create and build their own use cases for the technology.
Google Gemini Pro
Google says Gemini Pro, which will be free to use “within limits” via its new Google AI Studio building ecosystem, will also be “competitively priced” for those users looking for more functionality, offering a number of developer-friendly features.
Trained on Google’s fifth generation TPU v5p, Gemini Pro will launch with an initial 32k context window for text, with future versions set to expand on this, and include features such as function calling, embeddings, semantic retrieval and custom knowledge grounding and chat functionality.
The model will initially only accept text as input, and only generate text-based results, but Google has also produced a separate dedicated Gemini Pro Vision platform, which will also accept images as input.
For those looking to get their hands dirty, the company has released Google AI Studio, a free web-based development kit in your browser accessed via your Google Account.
Helpfully, Gemini-powered chatbots and search agents can be quickly built in a no or low-code environment.
Allowing users to quickly develop prompts, and receive an API key for app development, Google AI Studio can pull in text and images from your Google Drive account to help Gemini create results, such as writing a description of a house for a real estate agent.
Supporting 38 languages across over 180 countries at launch, Google AI Studio has a quota of 60 requests per minute, which Google says is 20x more than other free offerings and “suitable for most app development needs”.
Once complete, users simply click on a “get code” button to transfer their work to the IDE of their choice (including JavaScript, Python, cURL and more), use a number of quickstart templates on offer, or transfer it straight into Vertex AI.
When it's time for a fully-fledged AI offering in your business, Google’s Vertex AI platform will now offer “broad and deep support for Gemini Pro”, Google Cloud CEO Thomas Kurian noted, alongside over 120 open-source and third-party AI models.
Users can move between AI Studio and Vertex AI, allowing for enhanced customization whilst allowing your business to maintain full control of its data, which can be used to tune and tweak Gemini so that it is the ideal fit for your business needs, whilst also keeping the information secure and private.
Vertex AI will also offer a new set of models, including Imagen 2, the latest version of Google’s image creation model, which can now provide photorealistic images, as well as text rending and even logo generation - ideal for advertising or marketing campaigns.
Gemini Pro is available to developers everywhere now, with Google adding that its even-more-powerful Gemini Ultra will be launching “early next year”.
More from TechRadar Pro
- We've also rounded up the best AI writers around today
- These are the best laptops for programming
- Lots more Duet AI could be coming to Google Workspace soon
