Wednesday, August 26, 2026 · 1 of 5
OpenAI's Jalapeño inference chip posted its first benchmarks at Hot Chips: 1.5–1.9× more throughput per kilowatt and 1.7–3.6× lower end-to-end latency than current state-of-the-art inference systems
— deploying in OpenAI's own fleet by end of 2026 (Aug 25)