Industry and Infrastructure · Products and Tools · Business
OpenAI在Hot Chips大会公布Jalapeño基准测试结果，性能超越现有推理处理器
OpenAI在Hot Chips公布Jalapeño芯片基准测试，在InferenceX上每用户Token数和每千瓦吞吐量均超越现有推理处理器，预计2026年底小批量部署。
阅读 TechCrunch 原文为什么重要
自研推理芯片若成功，将降低OpenAI对英伟达的依赖，并可能重塑AI推理硬件市场格局。
关键事实
事实 1
Jalapeño在InferenceX基准测试中，每用户Token数和每千瓦吞吐量均高于现有最先进推理处理器。
来源与依据
Tested on SemiAnalysis’ InferenceX benchmark, Jalapeño registered both more tokens per user and more throughput per kilowatt than the currently available state-of-the-art inference processors.
事实 2
Richard Ho估计Jalapeño将于2026年底以极小批量部署，更大规模部署在2027年。
来源与依据
Ho estimated that Jalapeño would deploy at the end of 2026 “in very small volumes,” with more significant deployment coming in 2027.
事实 3
Jalapeño由OpenAI与Broadcom紧密合作开发，OpenAI自有模型参与开发过程。
来源与依据
First announced last October, Jalapeño was developed by OpenAI in close collaboration with Broadcom, with OpenAI’s own models assisting in the development process.