ChatGPT News 11/18/2026 AI Rating: Medium

OpenAI Models Optimize Jalapeno Chip Circuits and RISC-V Core Designs

#OpenAI#chip design#Jalapeno#AMD#RISC-V#Synopsys

OpenAI reported progress in AI-assisted chip design.

Jalapeno (OpenAI’s first custom inference chip)

  • Model-proposed changes reduced standard-cell area by about 13% in the vector unit and about 10% in the matrix unit (team estimate).
  • A matrix-unit multiplier redesign used about 15% less cell area and cut latency from two clock cycles to one; an adder redesign cut area by about 15% at the same two-cycle latency. These were silicon proven during bring-up.

AMD

  • Up to 14% power, performance and area (PPA) improvement over heavily optimized baselines on AMD’s flagship GPU IP, which AMD described as roughly a year of progress in under 5 months. Some changes went into a recent design sent for manufacturing; AMD also verified single-digit percentage gains on blocks of its upcoming flagship CPU. Synopsys tools verified that each rewrite preserved behavior.

RISC-V core evaluations

  • Tested on PicoRV32 and Ibex with a budget of about 112,000 tokens per run. GPT-6.2 Astra had the highest average score on both; GPT-6.1 Sol beat GPT-5.6 Sol. On PicoRV32, Astra’s score was about twice GPT-6 Astra’s.
  • Designing a core from a written specification scored higher than improving existing RTL: near 112,000 tokens, GPT-6.2 Astra averaged about 0.7 on PicoRV32 and about 0.75 on Ibex (9,881 scored runs in total).

Next: OpenAI is partnering with Synopsys to train GPT-Synopsys, a model meant to use Synopsys EDA tools as expert chip designers do.