AIThe Registerabout 2 hours ago
OpenAI Jalapeño chip beats Nvidia in inference benchmarks
6 min read
OpenAI's Jalapeño AI accelerator, developed with Broadcom, outperformed Nvidia's Blackwell system on Semianalysis's InferenceX benchmark, delivering more tokens per user and throughput per kilowatt. The chip, designed with OpenAI's models, deploys in small volumes at end of 2026, with significant volume in 2027.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
