Skip to content
IFAI

Jalapeño

Product2 stories

  1. IndustryReach 62

    OpenAI's custom chip is here, and it's fast

    OpenAI just published the first benchmark results for Jalapeño, its custom inference chip built around real LLM workloads. Tested on models like DeepSeek R1 and Kimi K2.5, it delivered up to 1.9x more work per watt and up to 3.6x lower end-to-end latency than commercial systems. Why it matters: Built in just nine months with AI assistance, Jalapeño proves OpenAI's full-stack strategy to slash serving costs as agentic workflows multiply compute demand.

  2. IndustryReach 42

    OpenAI and Broadcom drop Jalapeño, custom silicon for LLM inference

    OpenAI and Broadcom just revealed Jalapeño, OpenAI's first custom chip purpose-built for LLM inference. Developed in just nine months with help from OpenAI's own models, the chip reportedly delivers substantially better performance per watt than current accelerators and is already running GPT-5.3-Codex-Spark workloads in labs. Why it matters: OpenAI is pushing its full-stack strategy down to silicon, with plans to deploy at gigawatt scale by late 2026.