OpenAI: GPT-4o Search Preview

The OpenAI GPT-4o Search Preview model, created on March 12, 2025, and designed for web search within Chat Completions, demonstrates exceptional performance across several key metrics. It consistently ranks among the fastest models and offers among the most competitive pricing, both achieving an "Infinityth percentile" across six benchmarks. Reliability is a significant strength, with a perfect 100% success rate across all benchmarks, indicating minimal technical failures and consistent response delivery. In terms of specific benchmark performance, the model shows strong capabilities in Ethics (99.0% accuracy, 50th percentile) and Mathematics (92.0% accuracy, 71st percentile), suggesting robust reasoning and problem-solving skills. Coding also performed well at 89.0% accuracy (68th percentile). General Knowledge and Email Classification achieved 96.0% accuracy, though their percentile rankings were lower (44th and 38th respectively), indicating a competitive landscape in these areas. A notable weakness is observed in Instruction Following, where the model scored 0.0% accuracy, suggesting a significant area for improvement in complex, multi-step directive execution. Cost performance was consistently competitive across all evaluated benchmarks.