A Harvard researcher who himself scored a perfect 180 on the LSAT has found that multiple AI models can now do the same.
The Law School Admission Test is, by design, one of the hardest standardized exams in the world. Taken by roughly 170,000 candidates annually, it remains the most important factor in law school admissions. A perfect 180 is exceptionally rare.
Artificial intelligence just got one.
According to new research by Bonmu Ku, a graduate student at Harvard, four out of eight frontier reasoning models achieved the maximum scaled score of 180 across officially licensed LSAT questions: OpenAI’s GPT-5, Anthropic’s Claude Opus 4, DeepSeek-R1, and Moonshot’s Kimi K2 Thinking. It is the first documented instance of any AI system reaching the ceiling of the exam under controlled, independently verified conditions.
In March 2023, OpenAI reported that GPT-4 had scored a 163, the 88th percentile, on a partial LSAT evaluation. At the time, that was considered a milestone. Three years later, the number looks quaint. The gap between the 88th percentile and a perfect score may seem narrow on paper, but on the LSAT’s compressed upper scale, it represents a vast leap in reasoning consistency.
The models that reached a perfect score belong to a new generation of reasoning models. Rather than predicting the next most likely word, they generate an internal chain of thought, deliberating step by step before committing to an answer. His ablation experiments showed that when this thinking capability was turned off, performance dropped significantly. On an exam that rewards careful reasoning, that difference is significant.
The implications extend well beyond the lab. The LSAT has been the dominant admissions metric for American law schools for over seven decades. A vast test-preparation industry has been built around it. And the exam is not losing relevance. Applications are surging, with the 2025–2026 cycle seeing a 33% increase in submissions and a 22% rise in test registrations. A high score remains the single most important factor in law school admissions. The exam matters more than ever.
But the ceiling it was designed to measure no longer belongs to humans alone. The LSAT was built to identify the sharpest legal minds in the country. That benchmark has been met by systems that have never set foot in a courtroom.
