OpenAI Reports First Results for Custom Inference Chip 'Jalapeño': High-Speed, High-Efficiency AI Inference

On August 25, 2026, OpenAI announced the first results for "Jalapeño," a custom inference chip developed in-house. Jalapeño is designed to perform inference tasks for modern AI models with greater speed and energy efficiency. According to OpenAI, Jalapeño delivers faster and more energy-efficient AI inference while providing high throughput and low latency for contemporary models. The company stated that this announcement represents the initial results for Jalapeño, showcasing industry-leading speed and efficiency.