The company says the new toolkit can reduce model development and parameter extraction time from weeks to hours, allowing semiconductor teams to accelerate PDK delivery and support design technology collaborative optimization (DTCO). This release forms part of Device Modeling MBP 2026.
The announcement comes as the semiconductor industry embraces increasingly complex technologies, including gate-all-around (GAA) transistors, wide bandgap materials such as gallium nitride (GaN) and silicon carbide (SiC), and advanced integration approaches such as chiplets and three-dimensional stacking. Although these architectures offer improved performance, they also introduce new challenges for device modeling and parameter extraction.
Keysight noted that traditional compact modeling workflows rely heavily on physically-based models and manual tuning, often requiring engineers to tune hundreds of interdependent parameters across multiple operating conditions. This process can take several weeks, and tight development timelines may still result in less than optimal accuracy.
The company says its new machine learning toolkit addresses these challenges by combining neural network architecture with machine learning-based optimization. This toolkit includes an ML optimizer, automated extraction flows, and support utilities that can reduce parameter extraction from 200+ manual steps to less than 10 automated steps. This approach aims to accelerate PDK delivery, automate DTCO workflows, and reduce time to market.
According to Keysight, the toolkit enables global optimization of more than 80 parameters in a single run, capturing secondary effects, temperature variations, and dynamic behavior across DC, RF, and large signal domains. Automated workflows integrate with existing device modeling platforms and support Python-based customization. The workflow is designed to scale across multiple technologies, including FinFET, GAA, GaN, SiC, and bipolar devices, allowing modeling approaches to be reused across different process nodes.
Commenting on the release, Nilesh Kamdar, General Manager, Keysight EDA, said, “AI/ML is fundamentally transforming traditional workflows and methodologies for compact modeling. With our new machine learning toolkit, we are enabling our customers to deliver more predictive, high-quality models in significantly less time, accelerating PDK development and keeping up with rapidly evolving semiconductor technology.”
In addition to the toolkit, Keysight also announced updates to several other device modeling products.
Device Modeling MQA 2026 introduces new rules related to OMI and MOSRA aging model quality assurance, and Device Modeling WaferPro 2025 adds remote control functionality to support low frequency noise testing.
The latest A-LFNA 2026 release also introduces new low frequency noise stress test capabilities, allowing for a more streamlined transition from stress testing to noise measurements.
