Google introduced Gemini Pro to Vertex AI, expanding its GenAI model family

TL;DR:

  • Google introduces Gemini Pro on Vertex AI, expanding its GenAI model family.
  • Gemini Pro offers a lightweight alternative to Gemini Ultra and is now accessible to the public.
  • The Gemini Pro API supports 38 languages and regions, chat functionality, and filtering.
  • Gemini Pro Vision allows text and imagery processing, addressing previous limitations.
  • Developers can customize Gemini Pro for specific contexts and integrate it with external APIs.
  • Citation checking enhances the model’s reliability and trustworthiness.
  • Google emphasizes control, moderation, and governance options for Gemini Pro.
  • Competitive pricing, with input priced at $0.00025 per character and output at $0.00005 per character.
  • Exciting future features include powering conversational agents and search summarization.

Main AI News:

In a groundbreaking move, Google has expanded its GenAI model family with the launch of Gemini Pro on Vertex AI, the fully managed AI development platform. Following the recent unveiling of Bard and the Pixel 8 Pro, Gemini Pro steps into the limelight, offering developers a lightweight alternative to the highly capable Gemini Ultra model. While Gemini Ultra remains in private preview for a select set of customers, Gemini Pro is now accessible to the public, thanks to the new Gemini Pro API.

The Gemini Pro API is designed to empower developers to create their own applications with ease. It supports a wide range of features, including chat functionality, filtering, and is available in 38 languages and regions, including Europe. This sophisticated natively multimodal model combines advanced reasoning and coding skills, making it a game-changer in the world of AI.

Gemini Pro API Unleashes Creative Potential

The Gemini Pro API, by default, accepts text as input and generates text as output, much like other generative text model APIs. In addition, Google is introducing Gemini Pro Vision, an exciting new feature that allows text and imagery processing, including photos and videos. This capability brings Gemini Pro closer to OpenAI’s GPT-4 with Vision model, addressing a previous limitation and expanding its potential applications.

Customization and Integration with Vertex AI

Within Vertex AI, developers have the flexibility to customize Gemini Pro for specific contexts and use cases. This includes leveraging fine-tuning tools similar to those available for other Vertex-hosted models, such as Google’s PaLM 2. Gemini Pro can also be integrated with external APIs, enhancing its capabilities and relevance through third-party data or data from the web and Google Search.

Fact-Checking with Gemini Pro

Gemini Pro now includes support for citation checking within Vertex AI, enhancing its reliability and trustworthiness. This feature highlights the sources of information used by Gemini Pro to generate responses, ensuring accurate and well-informed results.

Control, Moderation, and Governance

Google is putting an emphasis on control, moderation, and governance options for Gemini Pro, addressing concerns about the model’s strength and reliability. These features offer developers greater control and transparency over the AI’s behavior, making it an attractive choice for a wide range of applications.

Attractive Pricing and Discounts

Google is committed to attracting developers by offering competitive pricing for Gemini Pro on Vertex AI. Input will cost $0.00025 per character, while output will be priced at $0.00005 per character. These prices represent a significant reduction compared to Gemini Pro’s predecessor. Furthermore, Google is offering a limited-time opportunity for developers to try Gemini Pro for free on Vertex AI until early next year.

Exciting Future Features

Google is not stopping at Gemini Pro’s launch. Several exciting features are in the pipeline, including the use of Gemini Pro to power custom-built conversational voice and chat agents, as well as search summarization, recommendation, and answer generation features within Vertex AI. These features are expected to arrive in early 2024, offering even more possibilities for developers.

Conclusion:

Google’s introduction of Gemini Pro to Vertex AI signifies a major advancement in the AI development market. With its versatile capabilities, customization options, and attractive pricing, Gemini Pro is set to empower developers and drive innovation in AI applications, potentially reshaping the landscape of AI development platforms. This move underscores Google’s commitment to maintaining a competitive edge and expanding its presence in the AI market.

Source