ainewsblitz.com

Breaking

OpenAI says Jalapeño inference chip beats Nvidia GB200 and GB300 on SemiAnalysis tests

  • Infra & Chips
  • Foundation Models

OpenAI published the first measured results for Jalapeño, its Broadcom-built custom inference ASIC, saying the chip delivered both more AI work per watt and lower latency than the best posted Nvidia GB200 and GB300 scores on SemiAnalysis’s public InferenceX benchmark. Hardware vice president Richard Ho told reporters the design aims at the usual trade-off between throughput and interactivity, claiming “the best of both worlds.”

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.

$20
Read this article
$29/month
Unlimited — all 27,531 articles, the full archive, and comprehension quizzes
Save 72%
$98/year
≈ $8.17/month
Unlimited, billed once a year