Google's Gemini 4 Argon ranks first on LegalReasonBench, a benchmark of long-horizon legal reasoning tasks written by practicing attorneys and released by micro1. Paige Bailey, who is tied to Google and Gemini, credits the score to the Google DeepMind team's work on multi-step reasoning, with the top result at 50.8%.
Love seeing Gemini 4 Argon take the #1 spot on LegalReasonBench! ⚖️✨
Long-horizon, multi-step reasoning over complex domain-specific tasks is something the @GoogleDeepMind team has been pushing hard on.
Huge congrats to @micro1_ai and the practicing attorneys who put this benchmark together -- and at 50.8%, still plenty of exciting headroom for what’s next! 🚀
AI models can be as convincing as a lawyer, but can they complete the complex multi-step reasoning a real legal matter requires? Today we're excited to introduce LegalReasonBench, a new benchmark to explore this question and test models on realistic, long-horizon legal matters authored by practicing attorneys. More info in thread 🧵
Source: 👩💻 Paige Bailey · x.comPublished · added here
