Harvey, a popular legal AI startup and one of the early companies backed by the OpenAI Startup Fund, is now expanding its use of AI models by incorporating technology from Google and Anthropic. The announcement was made in a blog post on Tuesday, where Harvey confirmed it will now use a variety of leading foundation models to power its legal services—marking a significant shift from relying solely on OpenAI’s models.
This move is especially notable given Harvey’s strong ties to OpenAI. Backed in 2022 as one of the first four startups in the OpenAI Startup Fund, Harvey quickly grew in value, reaching a $3 billion valuation following a $300 million Series D round in February 2025. That round included participation from top investors such as Sequoia, Coatue, Kleiner Perkins, and the OpenAI Fund itself.
Interestingly, Google’s venture arm, GV, led Harvey’s $100 million Series C funding in 2024 and participated in the Series D as well. However, Harvey didn’t adopt Google’s models at the time. Now, after internal testing with a benchmark called BigLaw, Harvey found that multiple models—particularly those from Google and Anthropic—excelled in specific legal tasks, prompting the decision to diversify its AI sources.
According to Harvey, Google’s Gemini 2.5 Pro performs strongly in legal drafting but struggles with pre-trial tasks, while OpenAI’s o3 and Anthropic’s Claude 3.7 Sonnet are better suited for complex legal reasoning. Instead of developing its own models, Harvey aims to fine-tune these external models for specific legal workflows.
The company also plans to launch a public leaderboard showcasing how various AI models perform in legal tasks. These rankings will include in-depth insights from legal professionals rather than relying solely on numeric scores.
Despite the shift, Harvey affirms its continued collaboration with OpenAI.
