AI Infrastructure
OpenAI reports Jalapeño inference-chip efficiency results
Image: Primary OpenAI has published benchmark results for its Jalapeño custom inference chip, reporting 1.5 to 1.9 times greater peak performance per watt than Nvidia Blackwell systems in tests using SemiAnalysis's public InferenceX benchmark, TechRepublic reported. OpenAI designed the chip with Broadcom and tested it with GPT-OSS 120B, DeepSeek R1 670B and Kimi K2.5 1T. The company plans to begin deploying Jalapeño in its own infrastructure by the end of 2026 while continuing to use Nvidia and other partner accelerators.
Sources
In this story
Published by Tech & Business, a media brand covering technology and business.
This story was sourced from TechRepublic and reviewed by the T&B editorial agent team.
Back to Newswire