-+ 0.00%
-+ 0.00%
-+ 0.00%

OpenAI says Jalapeño inference chip lifts performance per watt up to 1.9x in early tests

PUBT·08/25/2026 14:29:17
Listen to the news
OpenAI says Jalapeño inference chip lifts performance per watt up to 1.9x in early tests
  • OpenAI reported first benchmarked results for Jalapeño, its custom inference chip, on Aug. 25, 2026.
  • Across GPT-OSS 120B, DeepSeek R1, Kimi K2.5, delivered 1.5-1.9x more work per watt at peak throughput.
  • End-to-end latency ran 1.7-3.6x lower; interactive workloads delivered 2.1-4.1x higher performance.
  • InferenceX results: GPT-OSS peak mixed TPS/kW 85,448 vs 44,960; latency 1.03s vs 1.80s; min TBT 0.69ms vs 1.87ms.
  • Planned deployment in internal infrastructure by year-end; tapeout reached in nine months; Gen 2 in development, Gen 3 taking shape.


Disclaimer: This news brief was created by Public Technologies (PUBT) using generative artificial intelligence. While PUBT strives to provide accurate and timely information, this AI-generated content is for informational purposes only and should not be interpreted as financial, investment, or legal advice. OpenAI Inc. published the original content used to generate this news brief on August 25, 2026, and is solely responsible for the information contained therein.