vertex-ai

Exploring the Latest LLM Developments on Vertex AI

Tal Peretz

— 2 min read
Exploring the Latest LLM Developments on Vertex AI

In the ever-evolving landscape of artificial intelligence, Vertex AI continues to be a frontrunner with its innovative language model offerings. Among the most noteworthy are the recent expansions and improvements to its lineup of Large Language Models (LLMs), providing developers and researchers with advanced tools for a variety of applications.

Introducing the Vertex AI LLM Collection

The Vertex AI platform now boasts several cutting-edge models, each catering to different needs and use cases. Here’s a quick overview of the latest additions:

  • Gemini Models: The Gemini series, including Gemini 1.0 Pro, Gemini 1.5 Pro, and Gemini 1.5 Flash, is designed for batch predictions and seamlessly handles text, code, images, and video. Gemini 1.0 Pro is generally available, offering robust capabilities for diverse data processing tasks.
  • Llama Models: The Llama 3.1 405B model is a powerful new addition, enabling synthetic data generation, model distillation, and multilingual translation. Its general availability on Vertex AI makes it a valuable asset for complex projects involving large datasets.
  • Mistral Models: The Mistral Large (24.11) model provides managed model solutions, streamlining the integration of advanced AI into your workflow.
  • Gemma Models: Gemma 2 2B, developed by Google Deepmind, offers enhanced GPU utilization and input token masking, thanks to recent updates in fine-tuning notebooks.
  • Whisper Models: Whisper large v3 and Whisper large v3 turbo models are now part of the Model Garden, with improved fine-tuning capabilities for better performance.
  • Anthropic Claude Sonnet 3.5 v2: This model is generally available and represents another step forward in achieving nuanced AI-driven communication.

Cost-Effective Solutions

Vertex AI's pricing remains competitive, with both input and output costs set at $0.15 per million tokens, and a maximum token limit of 128,000 per request. This cost structure ensures that organizations can leverage powerful AI without incurring prohibitive expenses.

Functionality and Flexibility

One of the standout features of these LLMs is their support for function calling, a capability that enhances their adaptability and integration into existing systems. This feature allows for more dynamic interactions, enabling developers to craft applications that are responsive and intelligent.

With these advancements, Vertex AI solidifies its position as a leader in providing flexible, powerful AI solutions. Whether you're looking to improve data analysis, enhance translation services, or develop innovative applications, the Vertex AI platform offers the tools you need to succeed.

Read more