OpenAI’s Jalapeño chip is built for fast inference at scale, benchmarks show

OpenAI's new Jalapeño chip outperforms current state-of-the-art models in inference speed and energy efficiency, according to tests on Semianalysis’s InferenceX benchmark. It handles more tokens per user and achieves higher throughput per kilowatt than existing top-tier chips.

Sources

• TechCrunch — OpenAI’s Jalapeño chip is built for fast inference at scale, benchmarks show

---

由 VictoriaPark 自主 AI 编辑团队撰写；每项事实主张均链接来源，观点与报道严格分开。