Harvey, the legal tech startup backed by the OpenAI Startup Fund, has crossed an outstanding milestone. It has recently partnered with Anthropic and Google. This partnership allows Harvey to utilize several different foundation models, allowing it to deliver a more robust product to the legal industry. Harvey started out as one of the first four startups funded by the OpenAI Startup Fund. Today, it ranks with a $3 billion valuation, a testament to its rapid ascent in the tech universe.
Harvey’s leadership team, headed by CEO Winston Weinberg, has emphasized the importance of employing a variety of AI models tailored for specific legal tasks. Their internal benchmarking revealed that OpenAI’s o3 model excels in pre-trial tasks, closely followed by Anthropic’s Claude 3.7 Sonnet. This conclusion highlights Harvey’s mission to maximize the productivity and potential of AI technology in the legal space.
Back in February, Harvey shared the good news that they had closed a phenomenal $300 million Series D funding round. Sequoia led the Series B, with significant participation from Coatue, Kleiner Perkins, the OpenAI Fund and GV. This new round of funding continues to cement Harvey’s financial position. It puts the company in a stronger position to continue improving its AI agents.
“We are incredibly fortunate to have OpenAI as an investor in Harvey and key collaborator in our product. We are energized to add to our options for customers as we continue to serve the needs of our customers globally,” said Winston Weinberg. His statement is a strong indicator of the company’s ambition to improve and expand the product line by pulling in advanced AI capabilities from external sources.
Harvey’s about to join the ranks of other model companies that will be releasing a public Competitive leaderboard for model benchmark performance. This project aims to foster common ground and transparency between AI models. It further supports the principles of competition, arming attorneys with the power to choose based on performance data and competition standards. “Top lawyers provide nuanced insights into model performance not captured by single-score benchmarks,” noted Harvey, highlighting the significance of comprehensive evaluations.
The addition of models from competitors is an indication of Harvey’s plan to further refine these foundation models that perform best, into a version tailored specifically for the legal market. Google’s corporate venture firm invested in Harvey, but the startup did not immediately adopt its AI models, indicating a cautious approach to integration. Instead, Harvey said he would use generative models from Google and Anthropic via Amazon’s cloud services.
Harvey is deeply committed to leveraging inclusive, diverse AI models. Our recent legal benchmarks found that many foundation models do better than their underlying systems on specific legal tasks. “In less than a year, seven models (including three non-OAI models) now outperform the originally benchmarked Harvey system on BigLaw Bench,” stated the company. In a rapidly advancing digital age, this chilling discovery continues to highlight the importance of ongoing oversight and modification.
Hold on though, Harvey is still growing and innovating. Furthermore, its collaboration with industry leaders such as Anthropic and Google are sure to increase its competitive edge in the burgeoning legal technology market. The company is completely focused on improving AI-driven solutions built exclusively for legal professionals. All of this plays into its continued cred as a leader in the cultish niche market.
