Hardware Ready Ready to Bench?
Google Cloud Model Garden. Google Cloud has announced the integration of OpenAI models into its Model Garden, a dedicated catalog of artificial intelligence models for businesses. The announcement was made by Michael Gerstenhaber, Vice President of Product Management for Gemini at Google Cloud.
The available models include GPT-Astra and other OpenAI models, currently offered as a private preview to a select number of enterprise clients. You can express your interest through a form available on the official Google Cloud blog post.
Google Cloud Model Garden: why it matters
You will be able to access the models through Google Cloud APIs and integrate Codex CLI and the desktop app – OpenAI’s tools for assisted programming – with the Gemini platform for inference. In the future, the models will also be made available on Gemini for business and within Google Antigravity, Google’s AI development environment.
From a management perspective, the purchase of OpenAI models occurs directly through the Google Cloud Marketplace, with unified billing. This allows usage to be included in the amount of already subscribed spending commitments and to benefit from Flexible Savings Plans, Google Cloud’s flexible savings plans.
Google emphasizes its commitment to security, promising high standards. Currently, two features are under development and do not yet have a precise availability date: Zero Data Retention (ZDR), which guarantees the absence of data retention by the model provider, and Confidential Compute, technology for protecting data during processing. These characteristics are particularly relevant for companies with stringent compliance requirements.
This move follows six months after the revision of the agreement between Microsoft and OpenAI, which made Microsoft’s license to OpenAI’s intellectual property non-exclusive until 2032. On April 28, OpenAI models were introduced on Amazon Bedrock, the AWS service, in a limited preview.
What changes and what are the effects
Google states that it is the only hyperscaler to offer, on the same infrastructure and with a single contract, the three main model families – Claude, GPT, and Gemini – along with approximately 200 proprietary and open-weight models within the Model Garden and Vertex AI. It is important to note that in Amazon’s equivalent offering, Google’s Gemini is missing; AWS already offers OpenAI and Anthropic models on Bedrock.
According to Google Cloud, this offer simplifies contract management and the use of different language models for businesses, providing a complete and integrated ecosystem for developing artificial intelligence-based applications.
Source and further reading on Google Cloud Model Garden: original article.
* Content created with the assistance of artificial intelligence systems.