Anthropic, Google score win by nabbing OpenAI-backed Harvey as a user

Popular Legal AI Tool Harvey will now use the leading basic models from Anthropic and Google, Openai, Harvey’i strictly used declared In a blog post on Tuesday.

It is noteworthy because Harvey is one of the most successful real-supported portfolio companies in the Openai Starting Fund. Openai Start Fund is one Openai related stock Companies developing products above AI technologies, first of all, Openai is unique. When Harvey says he will not leave Openai, it’s just more model and cloud adding, it is still a big blow for Openai’s big rivals.

Harvey is one of the first four beginnings supported by Openai Start Fund In December 2022. When did this return Openai CEO Sam Altman He still managed the fund. (The others in this first cohort are descript, memorical and.)

Since then, Harvey, who has been grown mad, is now a $ 3 billion assessment start, in February, when declared By Sequoia, other major names similar to the Kleiner Perkins and Openai Foundation, Sequoia’s $ 300 million serial D.

Interestingly, Google’s enterprise lever, GV, LED Harvey’s $ 100 million series c In July 2024 (and Openai Foundation also took part in this period). However, Harvey immediately did not accept Google’s AI models after leaving Google’s AI models at the hat table immediately. (GV also took part in Harvey’s Series.)

So what does Harvey now believe in moving outside Openai models? Benchmark, Biglaw, a wide range of basic models in the startup internal developed internal, and some are better than others than others.

Instead of making efforts to make efforts, Harvey Fignation can embrace Obstetrics from other vendors (such as Google and Antropic “Amazon and Anthropik.

Using various models will help Harvey AI agents, the company says.

“In less than a year, seven models (including three non-OAI models) are superior to the Benchmarked Harvey system first,” Harvey wrote in the blog post.

Harvey’s Benchmark showed that a different basic models are better in special legal positions than others. For example, Google’s twins 2.5 Pro “Exceps”, but “Complex Cathed Rules”, but “Combat” is called “fight”, but “fight”.

According to Harvey, Openai’s O3 performs so many trials as Harvey’s tested.

Harvey also says it will now join the blog post Increased ranks of those who share a state leader board Model benchmark performance. The Board will sort out how basic thinking models do with legal positions. And the company will not only reduce the rankings, but also the best lawyers will provide Nuhance information in the model performance that is not seized in single-scale criteria. “

Thus, Harvey, who supported Openai, accepting the models of competitors, increases the pressure of those who support themselves to prove themselves (including Google). Openai should be very bothered in this account. While the Ei benchmarking increasingly complex and slightly politicalThis is a world that Openai is still shining.

“We are incredibly lucky to Openai, like an investor of Harvey and the main employee of our product,” said Harvey CEO Winston Weinberg Techcrunch. “And because we continue to serve the needs of our customers on a global scale, we get energy to add to our options for customers.”