Question
MATLAB R2025a, released in March 2025, marks a transformative update with hundreds of new features, focusing on a modernized desktop, enhanced graphics, AI integration, and workflow improvements across toolboxes.
Related Questions
Expert Answer
John Williams
PhD Expert
Answered Jul 21, 2026
MATLAB R2025a, released in March 2025, marks a transformative update with hundreds of new features, focusing on a modernized desktop, enhanced graphics, AI integration, and workflow improvements across toolboxes. It's the first release to fully leverage a WebGL-based graphics system (replacing Java), enabling better cross-platform consistency, including with MATLAB Online. R2025b, released in September 2025, builds on this foundation but emphasizes stability—introducing no major new features, only bug fixes and refinements based on user feedback from R2025a. Below, I'll break down the key highlights, drawing from MathWorks' official release notes.Major Desktop and User Interface Enhancements (R2025a)The desktop redesign is the star of R2025a, making MATLAB more intuitive and customizable:
Simulink Updates (R2025a)
- Customizable Sidebars and Panels: New sidebars (left, right, bottom) for quick access to tools like Debugger, Project, Source Control, Test Browser, Code Issues, and Code Outline. Tailor layouts for specific workflows—e.g., dock the Debugger for troubleshooting without cluttering your view.
- Themes and Accessibility: Switch between light/dark themes for the desktop, figures, and apps (e.g., set(0, 'DefaultUIBackgroundColor', [0.15 0.15 0.15]) for dark mode). Enhanced screen reader support for Command Window navigation and code editing improves inclusivity.
- Figure Container: Figures now open as tabs in a single container by default, supporting tiling, resizing, grouping, and annotations via an updated toolstrip—ideal for comparing plots without juggling windows.
- Desktop Search and Feedback: Search launches apps, settings, help, examples, or MATLAB Answers directly. A new toolstrip Feedback button lets you report issues or suggest ideas seamlessly.
- Live Editor Upgrades: Save live scripts/functions in a plain-text .mlx format for better source control integration (e.g., Git). Embed YouTube videos and export to Markdown for sharing.
- Graphics Modernization: WebGL rendering for smoother, theme-aware plots with automatic color adaptation. App Designer gets a richer, scalable UI with more components—now the go-to for building apps.
- MATLAB Copilot: A groundbreaking AI assistant (new product) that integrates generative models into workflows for code suggestions, debugging, and analysis. Access via the desktop for tasks like auto-completing functions or explaining errors.
- Live Code Files: Full support for plain-text .m live scripts, enhancing version control and collaboration.
- Variables Editor: View sparklines and summary stats for tables/timetables; create linked plots that sync with workspace changes.
- Code Optimization: Automatic parallelization in MATLAB Coder for faster generated code. New Python Package Compiler App simplifies packaging MATLAB code for Python ecosystems.
|
Toolbox
|
Key New Features
|
|---|---|
|
Deep Learning Toolbox
|
Improved PyTorch model import/performance; faster training with AI-based optimizations.
|
|
Phased Array System Toolbox
|
RIS modeling with controllable elements; bistatic localization via TOA/TDOA; cascaded RF impairment models.
|
|
Radar Toolbox
|
Parallel workflows for bistatic/multistatic radar simulations.
|
|
Sensor Fusion and Tracking Toolbox
|
Tracking Data Importer app; easier multi-object tracker tuning with truth data.
|
|
Antenna Toolbox
|
AI-optimized antennas via SADEA API; scalar port/field analyses.
|
|
RoadRunner & Scenario
|
Programmatic APIs for road scenes/automotive scenarios; snappable templates and elevated features like tunnels.
|
|
Risk Management Toolbox
|
Credit model validation; ES backtesting for VaR/ES models.
|
|
HDL Verifier
|
Cosimulation with Synopsys VCS simulator.
|
- Layout-as-You-Go: Auto-straighten signal lines during block moves/resizes.
- Model Referencing: Local solvers (e.g., ode23t) for referenced models, enabling faster top-model steps while maintaining accuracy.
- Fast Restart: Quicker simulations in Rapid Accelerator mode, plus Python Code Blocks for direct integration.
- Projects UI: Simplified interface for source control and dependency analysis.
Have a different question? Ask here