Developers and businesses, embrace Gemini Pro now

Introducing Gemini Pro: AI Model for Developers

Last week marked the launch of Gemini, our most advanced AI model, with Ultra, Pro, and Nano options. Gemini is now integrated into products, such as Android’s Pixel 8 Pro and a tailored version in Bard.

Today, Gemini Pro is open for developers and enterprises, ready for customization based on your needs. Ongoing improvements will be made based on user feedback.

Key Features:

• Outperforms similar-sized models in research benchmarks.
• 32K context window for text, with plans for expansion.
• Free usage, competitively priced in the future.
• Supports 38 languages worldwide.
• Features include function calling, embeddings, semantic retrieval, custom knowledge grounding, and chat functionality.
• Text and multimodal (text and imagery) outputs available.

SDKs support Gemini Pro in Python, Android (Kotlin), Node.js, Swift, and JavaScript.

Google AI Studio:

• Free tool for prompt development and API key acquisition.
• Sign in with your Google account.
• 60 requests per minute, 20x more than other free options.
• API and Studio data may be accessible to reviewers for quality improvement.

Transition to Vertex AI on Google Cloud:

• Fully-managed AI platform with customization options.
• Tune and distill Gemini with your company’s data.
• Build search and conversational agents with low code/no code.
• Deploy confidently with data and IP ownership.

Vertex AI Access:

• Try Gemini models at no cost until general availability.
• Pricing per 1,000 characters or per image after general availability.

For more information, visit the Google Cloud blog.

Anticipating the Future

We are thrilled to announce that Gemini is now accessible to developers and enterprises. As we continually refine this platform, your input will be instrumental in enhancing its performance. To explore more and commence your Gemini projects, visit ai.google.dev, or leverage the robust capabilities of Vertex AI with enterprise-grade controls on your own data.

In the early months of the upcoming year, we are set to unveil Gemini Ultra – our most extensive and proficient model designed for highly intricate tasks. This release follows additional fine-tuning, rigorous safety testing, and the invaluable feedback received from our partners. Furthermore, we plan to extend Gemini’s integration to more developer platforms, including Chrome and Firebase.

We are eager to witness the innovative solutions you create with Gemini.