OpenAI's First Custom Inference Chip Jalapeño Beats Commercial Systems
firstadopter · x · 2026-08-25
OpenAI released the first measured performance results for Jalapeño, its first custom inference chip:
Performance:
- Delivers higher peak throughput per kilowatt and lower token latency than commercial systems on the InferenceX benchmark (GPT-OSS 120B).
- Shows strong performance on DeepSeek R1 and Kimi K2, extending gains across model families.
Development & Efficiency:
- AI-Assisted Design: AI enabled the team to move from initial design to tapeout in nine months, shortening verification loops.
- Circuit Optimization: AI helped optimize arithmetic circuits to fit more compute performance.
- Jevons Paradox: Greater efficiency makes more uses worthwhile, expanding consumption and economic activity.
Roadmap:
- Deployment planned within OpenAI's infrastructure by year-end.
- Gen 2 is deep in development, and Gen 3 is taking shape.
More from Infra
- Actual Computer serves 8,000+ tokens/s from a single NVIDIA DGX Station — cephaloform · 2026-10-16
- Jensen Huang joins BG2 podcast: does AI revenue justify capex, '27 bottlenecks — elonmusk · 2026-10-16
- Mistral's on-device infra rack revealed: DGX Spark, Mac Studio and more — gregmushen · 2026-10-16
- NVIDIA and Sweden's Wallenberg launch Sferical One, a shared AI supercomputer — nvidia · 2026-10-16
- Next infra frontier: self-evolving systems that adapt and self-verify — _shreya_s · 2026-10-16
- Xiaomi's MiMo RL dashboard reveals $3.47M spent on just 30 training steps — bycloud · 2026-10-16