Harvey hires for its Private Model Program

On September 24, Gabe Pereyra said Harvey was hiring for its Private Model Program after completing the first part of a model strategy he had described as building Harvey’s own model and using it to help customers do the same. Pereyra said the company was seeing huge demand from law.

Pereyra had said on August 20 that Harvey had post-trained a model called Tenet. He said Tenet achieved SOTA on LAB, generalized 3rd party legal benchmarks and used sub-agents for domain specific capabilities. He also said Tenet used Kimi K3 as its base and was post-trained in collaboration.

Winston Weinberg said in July that Harvey was training a series of legal foundation models, launching Harvey Research and investing heavily in model training. He also described work involving 100M-token data rooms and synthetic legal matters for model training. Harvey later described itself as becoming a full-stack AI company, scaling data with mercor and helping law firms build and own their intelligence.