OpenAI models find up to 14% PPA gains on AMD GPU IP, improve custom Jalapeño chip

OpenAI · x · 2026-11-19

OpenAI reports that its models, working with AMD, identified up to 14% PPA improvement on flagship GPU IP in under five months—building on designs AMD engineers had optimized for years. The same approach improved Jalapeño, OpenAI's first custom inference chip, with gains verified on real silicon. Charts show newer models (GPT-6.2 Astra and others) far outperforming GPT-5.5 on RTL optimization and spec-to-RTL tasks for Ibex and PicoRV32. OpenAI is also partnering with Synopsys to bring its model capabilities into Synopsys's EDA design and validation tools.

Related event: OpenAI Models Boost AMD Flagship GPU Design by Up to 14% PPA(2 posts)→

Original post →

More from Infra

Infra channel →