OpenAI's o3 AI model scores lower on a benchmark than the company initially implied | TechCrunch
techcrunch.com
A discrepancy between first- and third-party benchmark results for OpenAI's o3 AI model is raising questions about the company's transparency and model testing practices.