What’s new in STM32CubeMX

[Maxime_MARCHETTO1]

STMicroelectronics has released new versions in STM32Cube ecosystem, expanding STM32CubeMX support to more STM32 products and enabling configuration of innovative on‑chip IP for smarter, more efficient embedded designs.

The new release brings STM32CubeMX support for the latest STM32H5, STM32U3, and STM32WBA2 devices, helping developers quickly start new designs and reuse existing workflows across a growing STM32 portfolio.

STM32Cube v1.22 also adds configuration support for two advanced peripherals on selected STM32 devices:

• PLAy (Programmable Logic Array) provides a small, configurable logic fabric on‑chip, ideal for custom glue logic, simple state machines, and deterministic event detection. By wiring internal events and external GPIO signals directly into PLAy, applications can react in hardware with cycle‑accurate timing, offloading the CPU and often eliminating external logic components.
  
  

The PLAy is supported on latest STM32H5E4/5F4 and STM32H5E5/5F5 product lines

• HSP (Hardware Signal Processing) targets real‑time signal processing at very low power. It offloads filtering, spectral analysis, and feature extraction for continuous sensor streams (such as vibration, audio, or current sensing) to a dedicated hardware engine, allowing the main core to stay in low‑power modes until meaningful events are detected.
  
  

HSP is now available on new STM32U3B5/3C5 product lines.

Together, PLAy and HSP enable more responsive, efficient, and compact embedded systems by combining fast, deterministic control logic with rich, pre‑processed sensor information.

On the MPU side, STM32CubeMX now supports new configuration (M33-TD) of STM32MP2 devices, enabling developers to manage the Arm Cortex‑A35 as coprocessor from the Arm Cortex‑M33 core (booting on Cortex-M33 Trusted Domain flavor) on both evaluation and discovery boards. The release integrates with the newly available OpenSTLinux 6.2, which adds key capabilities such as STM32MP15 PSCI OSI, STM32MP2 A35‑TD PCIe Endpoint support, and GPU OPP management, and completes the Cortex‑M33 Trusted Domain flavor for mass‑market applications (including secure boot, system reset requests from either Cortex‑A35 or M33, and early panel splash screen driven by a non‑secure Cortex‑M33 application).

Additional resources

• Download STM32CubeMX
• Video: [Getting started] STM32CubeMX
• Download OpenSTLinux
• Release note: STM32 MPU OpenSTLinux release note - v6.2.0 - stm32mpu

First published on Mar 23, 2026

[Ahmet Yasin CİVAN]

This is a fantastic update! Expanding support to the new STM32H5, STM32U3, and STM32WBA2 devices keeps the ecosystem strong. I am particularly excited about the configuration support for advanced peripherals like PLAy and HSP. Offloading tasks from the CPU using programmable logic and hardware signal processing is exactly what modern embedded systems need for better efficiency and low power.

However, as ST continues to introduce these highly advanced and complex on-chip IPs, the configuration process naturally becomes steeper for developers. This brings me to a suggestion: Native AI (LLM) support directly integrated into STM32CubeMX.

We are living in the AI era, and ST has already shown initiative with the STM32 Sidekick. Bringing a native LLM assistant into the core CubeMX workflow could be a game-changer, especially for configuring these new, complex peripherals.

Here are a few ways an integrated AI assistant could elevate STM32CubeMX:

1. Prompt-to-Config for Advanced Peripherals: Configuring a logic fabric like PLAy or setting up filtering for HSP can be highly complex. Instead of manually mapping internal events and GPIOs, developers could simply type to an AI assistant: "Configure PLAy to act as a simple state machine that triggers an event when GPIO X goes high and GPIO Y goes low" or "Set up HSP for continuous vibration sensing on STM32U3, keeping the main core in sleep mode until a threshold is reached." The LLM could instantly translate this intent into the correct CubeMX graphical settings.

2. Smart Dependency and Conflict Resolution: As we add more coprocessor setups (like the M33-TD configuration for STM32MP2 mentioned in the update), managing shared resources and clock trees gets complicated. An integrated AI could instantly spot conflicts and suggest direct fixes: "Your current Cortex-M33 setup conflicts with the A35 peripheral assignment. Re-route peripheral Z to resolve this."

3. Context-Aware Code Snippets for New Features: Since PLAy and HSP are relatively new, developers will need help writing the firmware to utilize them. A built-in AI could generate tailored HAL code snippets or initialization sequences based exactly on the current GUI configuration, saving hours of reading through new application notes.

4. Dynamic Datasheet Querying: Instead of leaving the tool to read a 2000-page Reference Manual to understand the new registers for PLAy or HSP, developers should be able to ask the CubeMX AI: "Explain the timing constraints for wiring an external GPIO to the PLAy module" and get a precise, summarized answer.

With the addition of such powerful hardware features in this release, wrapping them in an intelligent, AI-driven configuration workflow would make STM32CubeMX unstoppable. I hope the ST development team considers integrating native LLM capabilities in future updates!