Release Notes
Multi-Model Support
Harvey now supports Claude and Gemini, giving teams access to multiple leading foundation models—with better default results and more control when needed.
Harvey has launched Multi-Model Support, enabling workspace admins to roll out Anthropic’s Claude and Google’s Gemini models directly via self-serve permissions. These models join our existing OpenAI-powered systems, offering teams greater flexibility and performance across legal tasks.
With multiple foundation models now available in Harvey, users benefit from:
- Better default results, as Harvey routes tasks to the most capable systems
- More control, through admin-managed model access and optional Model Selector
What’s new:
- Claude + Gemini models: Now available to be rolled out by workspace admins
- Model Selector (opt-in): Enables users to switch models in Assistant, Vault, and Workflow Builder
- Expanded support: OpenAI models remain fully supported
Secure and compliant:
Claude and Gemini are integrated via AWS Bedrock and Google Vertex AI, with the same privacy and security guarantees that underpin all of Harvey’s systems.
Why it matters:
This launch marks the first step in Harvey’s shift toward a more adaptive and extensible AI platform for professional & legal work. Multi-model support allows us to:
- Dynamically match users with the right model for the right task
- Customize workflows based on specific needs
- Accelerate evaluation and integration of future models offering unique advantages
Read the blog post: https://www.harvey.ai/blog/expanding-harveys-model-offerings