Harvey Expands AI Capabilities by Embracing Google and Anthropic Models

In a significant move within the legal tech industry, Harvey, the popular legal AI tool, has announced its decision to incorporate leading foundation models from both Google and Anthropic. This strategic shift, detailed in a blog post on Tuesday, marks a pivotal moment for the company, which has quickly become one of the standout entities backed by the OpenAI Startup Fund.

Harvey’s announcement is not just about diversification; it reflects a broader trend where AI startups are seeking flexibility and versatility in an increasingly competitive market. Although the company reassured that it will maintain its ties with OpenAI, the integration of additional models signals a shift towards a more inclusive approach that aims to harness the strengths of various AI platforms.

The Rise of Harvey and Its Connections

Founded as part of a select group of startups in December 2022 that received backing from the OpenAI Startup Fund, Harvey has experienced tremendous growth. With a valuation soaring to $3 billion after a successful $300 million Series D funding round led by Sequoia, the company has emerged as a powerful player in the legal technology landscape. Notably, prominent venture capital firms such as Coatue and Kleiner Perkins also participated in this funding, solidifying Harvey’s reputation.

Interestingly, Harvey had previously secured $100 million in funding from GV, the venture arm of Alphabet, in July 2024. Despite having such substantial backing from Google’s corporate venture firm, the company took its time before adopting Google’s AI models. This cautious approach is now yielding dividends as they broaden their capabilities.

Why the Shift? Insights from Internal Benchmarking

What prompted Harvey to transition beyond relying solely on OpenAI’s models? According to the startup, their internal benchmark, known as BigLaw, revealed that multiple foundation models are not only becoming more proficient in legal tasks but that specific models outperform others in certain areas.

Rather than developing models from the ground up, Harvey determined it would be more efficient to utilize high-performing models from various vendors, including Google and Anthropic. This strategy allows them to fine-tune existing AI to better serve the legal industry. Notably, Harvey has shown that their benchmark indicates that different models excel in different tasks, with some doing well in legal drafting while others perform better in pre-trial tasks.

For example, Harvey noted that Google’s Gemini 2.5 Pro is exceptional at drafting legal documents but falls short in pre-trial tasks due to challenges in grasping complex evidentiary rules. Conversely, OpenAI’s o3 model is effective at pre-trial activities, with Anthropic’s Claude 3.7 Sonnet close behind.

The Future of AI in Legal Tech

In an intriguing development, Harvey plans to join the ranks of companies providing a public leaderboard that benchmarks AI model performances on legal tasks. This initiative aims to go beyond simple numerical rankings by incorporating qualitative insights from top legal professionals on model performance. Such transparency could serve as a valuable resource for legal institutions navigating the rapidly changing landscape of AI.

This adoption of competitors’ models signifies Harvey’s commitment to delivering the best possible tools to their clients while fostering healthy competition among their backers. “We are incredibly fortunate to have OpenAI as an investor in Harvey and a key collaborator in our product,” stated Harvey CEO Winston Weinberg in an interview with TechCrunch. “And, we are energized to add to our options for customers as we continue to serve the needs of our clients globally.”

The landscape of legal tech is evolving as AI capabilities expand and improve. With competitors such as Google and Anthropic stepping into the fray, Harvey’s adaptive approach positions it well to stay ahead of the curve. As legal tasks become increasingly complex and varied, the ability to leverage the best models on the market will be crucial for any AI-driven legal platform.

In summary, Harvey’s pivot to incorporate multiple AI models not only enhances its product offerings but also reflects a broader trend within the industry towards collaboration and competition. This shift highlights the importance of flexibility in the rapidly changing technology landscape, suggesting we can expect more innovative solutions in the legal sector in the near future.

For more insights into the evolving role of AI in various sectors, explore our articles on Artificial Intelligence Applications, Legal Tech Innovations, and AI Model Benchmarking.