Release Notes

Fable 5.1 is now available as a model provider in Harvey, giving workspaces that opt in access through the model selector.

GPT-6 Astra is now available as a model provider in Harvey, giving workspaces that request it access through the model selector.

Harvey now supports Mistral as a model provider, giving teams access to Mistral Medium 3.5.

Fable 5 is now available as a model provider in Harvey, giving workspaces that request it access through the model selector.

Opus 5 is now available in Harvey’s Model Selector across Assistant, Vault, and Workflow Builder.

GPT-5.6 Sol is now available in Harvey’s Model Selector across Assistant, Vault, and Workflow Builder.

Claude Sonnet 5 is now available in Harvey's Model Selector across Assistant, Vault, and Workflow Builder.

Opus 4.8 is now available in Harvey’s Model Selector across Assistant, Vault, and Workflow Builder.

Opus 4.7 and GPT-5.5 are now available in Harvey’s Model Selector across Assistant, Vault, and Workflow Builder.

GPT-5.4 is now available in Harvey’s Model Selector across Assistant, Vault, and Workflow Builder.

Opus 4.6 and Sonnet 4.6 are now available in Harvey’s Model Selector across Assistant, Vault, and Workflow Builder.

GPT-5.2 is now available in Harvey’s Model Selector across Assistant, Vault, and Workflow Builder.

GPT-5.1 is available in Harvey’s Model Selector across Assistant, Vault, and Workflow Builder.

Harvey Assistant just got smarter—GPT-5 is now the default model for all chat and draft responses, bringing major upgrades in reasoning, reliability, and writing quality.

Harvey now supports Claude Sonnet 4.5, Anthropic’s latest reasoning model—optimized for code, numbers, and complex legal work.

Use GPT-5 in Assistant, Vault, and Workflow Builder—unlocked via the model selector for faster reasoning, smarter drafting, and advanced multi-step orchestration.

Harvey now supports Claude and Gemini, giving teams access to multiple leading foundation models—with better default results and more control when needed.

Assistant now runs on GPT-4.1, delivering better output quality, longer responses, and an updated knowledge cutoff.

A new benchmark measures how well LLMs perform on real-world BigLaw tasks—Harvey leads in producing expert-level outputs.