GPT-5 Outperforms Human Judges with 100% Legal Compliance in Landmark Study
Research reveals GPT-5 achieved 100% legal accuracy vs 52% for human judges, raising questions about AI's role in judicial decision-making.
Research reveals GPT-5 achieved 100% legal accuracy vs 52% for human judges, raising questions about AI's role in judicial decision-making.
OpenAI announced OpenAI for Science, a dedicated team leveraging GPT-5 to help scientists find references, sketch proofs, and test hypotheses. GPT-5.2 scores 92% on PhD-level knowledge benchmarks, positioning AI as a research collaborator.