Anthropic & Google Land OpenAI-Backed Harvey as Client
▼ Summary
– Harvey, a legal AI tool, will now use foundation models from Anthropic and Google alongside OpenAI’s, expanding its AI provider options.
– Despite being backed by the OpenAI Startup Fund, Harvey’s shift to competitors’ models is a significant win for Anthropic and Google.
– Harvey’s $3 billion valuation and rapid growth were fueled by major funding rounds, including a $300 million Series D led by Sequoia.
– The decision to diversify models was driven by Harvey’s internal benchmark, BigLaw, which showed varying strengths in different legal tasks across AI models.
– Harvey plans to publish a public leaderboard ranking AI models on legal tasks, adding pressure on providers like OpenAI and Google to improve performance.
Harvey, the high-profile legal AI startup valued at $3 billion, is expanding its technology partnerships beyond OpenAI to include foundation models from Anthropic and Google. This strategic shift comes despite Harvey being one of the earliest beneficiaries of OpenAI’s startup funding program.
The move signals a broader industry trend where AI companies increasingly adopt a multi-model approach rather than relying on a single provider. Harvey’s decision follows rigorous internal testing using its proprietary BigLaw benchmark, which revealed that different models excel at distinct legal tasks. For example, Google’s Gemini 2.5 Pro performs exceptionally well in legal drafting but falters in pre-trial applications, while OpenAI’s models and Anthropic’s Claude 3.7 Sonnet demonstrate stronger capabilities in courtroom-related workflows.
Harvey’s rapid growth—backed by top-tier investors like Sequoia, Coatue, and Kleiner Perkins—has positioned it as a leader in legal AI. The startup’s recent $300 million Series D round underscores investor confidence in its vision. Notably, Google’s venture arm, GV, participated in Harvey’s funding rounds, though the company only now integrates Google’s AI models into its platform.
Beyond diversifying its model providers, Harvey plans to launch a public leaderboard tracking how major AI systems perform on legal tasks. Unlike traditional benchmarks, this ranking will incorporate detailed evaluations from practicing lawyers, offering deeper insights into real-world applicability.
While OpenAI remains a key investor and collaborator, Harvey’s embrace of competing models highlights the intensifying race for AI dominance. As the legal tech space evolves, Harvey’s multi-vendor strategy could set a precedent for other AI-driven enterprises seeking flexibility and performance optimization.
Harvey CEO Winston Weinberg emphasized the company’s commitment to customer needs, stating, “We are energized to expand our options while continuing to leverage OpenAI’s groundbreaking work.” With AI benchmarking becoming more nuanced, Harvey’s approach ensures it stays at the forefront of innovation while keeping pressure on its partners to deliver cutting-edge solutions.
(Source: TechCrunch)
