Harvey's legal agent benchmark grew to 1,671 tasks and still publishes no model scores three months on.

harveyai/harvey-labs is trending again (692 stars, +47 on August 9). The README now states 1,671 tasks across 24+ legal practice areas plus contracting, up from 1,250 at the May 6 launch, which also cited 75,000+ expert-written rubric criteria (GitHub). Harvey said at launch that "initial results on LAB for leading open and closed-source models will be published in the coming weeks." Three months later the repo ships the dataset, rubrics, and execution framework with no comparative scores. If you're evaluating agents for regulated professional work, that absence is the finding. Run it yourself; the framework is there.