R2025a Pre-release highlights
Walter Roberson
on 17 Jan 2025
Latest activity Reply by William Thielicke
on 14 Mar 2025 at 19:35
This topic is for discussing highlights to the current R2025a Pre-release.
69 Comments
Sign in to participate
Hi, is it a known issue that the GUI is much slower than in e.g. 2024b? Will Mathworks work on this? Do I need to supply a screen capture with a comparison to show the difference?
I noticed that inside Source Control "Rename" has disappeared in 2025A.
Someone has an explanation?
Hi Marco
I've just taken a look and can confirm that its disappeared. I don't know why and am investigating internally. Will update when I have something.
Cheers,
Mike
The team have got back to me on this. Git integration has been improved in this release so all you need to do is rename the file in the Files Panel, and the source control integration will do the right thing under the hood.
Give it a try and let me know how it goes.
Another important issue is the different unexpected behaviour in R2025A in annotation.
For example in R2010a-R2024b the code below
h=figure;
fs=14;
dim = [.02 .830 0.1 0.1];
str='$\begin{array}{rrrrrrrr} \\ $\boldmath{$i$}$ & $\boldmath{$x_i$}$ & $\boldmath{$y_i$}$ & $\boldmath{$x_i^2$}$ & $\boldmath{$x_iy_i$}$ & $\boldmath{$\hat y_i$}$ \\ 1 & 1330 & 120 & 1768900 & 159600 & 78.9443 \\ 13 & 1380 & 260 & 1904400 & 358800 & 85.4261 \\ \bf - & \bf 21330 & \bf 1550 & \bf 37505100 & \bf 2868250 & \bf 1550 \\ \end{array}$';
annotation(h,'textbox',dim,'FitBoxToText','on','String',str,'Interpreter','latex','FontSize',fs);
produces
On the other hand in R2025A it produces
I already showed this issue to the support but with I did not have feedback.
Good day everyone, is there a way to see function help directly in the app instead of opening a browser page like before?
Thank you all fr your answers
Hi Leonardo
There is no way to see the full documentation in the app since the help browser has been removed. It is, however, still possible to get the text-only help from the command line. E.g. to see the help for the ones function
help ones
Hi Leonardo,
if after highlighting the function you click on F1 the pop up still appears and Learn More (see screenshot below) leads you directly to the function
Similarly, as suggested by Mike, if you type help FunctionName
if at the end you click on "Documentation for ones"
you go directly to the Docuementation page.
Given that this is a brain storming page why after F1 on the function you have "Learn More" and after help xxx you have "Documentation for xxx"?
In our FSDA toolbox after pressing the F1 key we have at the end Link
"Link to the help page for this function"
Same thing through the Command Line if you type help FSR.
A bug that I have found (at least in my computer) is that for any function belonging to MATLAB toolboxes if you pass through the F1, the link does not work.
In other words if I type
help linprog (function from optimization toolbox) and at the end I click on
Documentation for linprog
it works
on the other hand if I click on F1 after selecting the name of the function
I get
and doc linprog is not clickable. In summary there is a difference between MATLAB help functions and, functions from MATLAB toolboxes.
Thanks for the careful and detailed description. I was not able to reproduce that error, so I wanted to check on the steps so I can report it if it is a bug. I clicked f1 on the word 'linprog' in a script. I get the attached image, which does not have the link for 'documentation for linprog' at the botton (even when I scroll down), but it does have a Learn More link that takes me to the appropriate doc. I do see 'documentation for linprog' when i use 'help linprog' but that link works fine for me as well. Am I following your steps correctly? I am using 25b prerelease update 1. Can you make sure you have the latest update and try again? Also have you set the doc to not be installed?
I can confirm that now after the 25b prerelease update 1 I see what you see and there is Learn More at the end which brings me to the help of the function.
The only one minor "inconsistent thing" which remains is that if you clik on help linprogr in the Command Line you then have to click on
"Documentation for linprog"
On the other hand if you start from F1, you have to click on
"Learn More".
That said, now with update 1 in both ways you go directly to the documentation for that particular function.
Thanks for the confirmation and description, ill report the inconsistency as a possible enhancement in our tracking system
Is it intentional that the List box UI component cannot be clicked+drag anymore to select multiple items? I am pretty sure that the younger generation isn't aware that they can use shift and ctrl to select multiple items....
I am not sure if it's intentional but I confirm that I get the same behaviour. For the avoidance of doubt, are you referring to something like this
lb = uicontrol('Style','listbox','String',{'1','2','3'},'Value',2,'Max',3,'Min',1,'Callback',@(h,evt) disp(h.Value),'Position',[10 10 100 100]);
in R2024b I can click and drag to select multiple items and so do shift+CTRL. In R2025a, I can only use Shift+CTRL.
I'll get this added to the database and see what comes back from dev.
Yes, this is what I meant!
It is now in the database as an enhancement request. I don't have an ETA on when it will get implemented though.
I am very positive about the changes, looks like a really good thing. But I guess I have to spend some time to get used to it and discover all the new features. I guess that you will make a "transition" tutorial video for users of the old UI?
What I don't like too much, is that developers of UI tools on FEX were not "warned" or somehow involved a bit ealier about this major change. Some popular toolboxes and functions from FEX completely break with the new Matlab UI. My own toolbox also requires a lot of extra work. It would have been great if I had more time to fix all the new bugs. There are probably 6 weeks left, and I think I can make it, but I don't know about others that have more severe problems than me.
Thanks for the feedback William. I'll make sure it gets heard internally.
nebula
ans =
0 0.54336 0.98326
0 0.54129 0.98276
0 0.53921 0.98223
0 0.53712 0.98167
0 0.53503 0.98109
0 0.53292 0.98047
0 0.5308 0.97982
0 0.52868 0.97915
0 0.52654 0.97844
0 0.5244 0.9777
0 0.52224 0.97694
0 0.52008 0.97614
0 0.5179 0.97532
0 0.51572 0.97446
0 0.51352 0.97357
0 0.51132 0.97266
0 0.5091 0.97171
0 0.50687 0.97073
0 0.50464 0.96972
0 0.50239 0.96868
0 0.50013 0.96761
0 0.49786 0.96651
0 0.49558 0.96538
0 0.49328 0.96422
0 0.49098 0.96303
0 0.48867 0.96181
0 0.48634 0.96055
0 0.484 0.95927
0 0.48165 0.95795
0 0.47929 0.9566
0 0.47692 0.95522
0 0.47453 0.95382
0 0.47213 0.95237
0 0.46972 0.9509
0 0.4673 0.9494
0 0.46487 0.94787
0 0.46242 0.9463
0 0.45996 0.94471
0 0.45749 0.94308
0 0.455 0.94142
0 0.4525 0.93973
0 0.44999 0.93801
0 0.44747 0.93626
0 0.44493 0.93448
0 0.44238 0.93266
0 0.43981 0.93082
0 0.43723 0.92894
0.022235 0.43463 0.92704
0.067379 0.43203 0.9251
0.09768 0.4294 0.92313
0.12138 0.42676 0.92113
0.14137 0.42411 0.9191
0.15893 0.42144 0.91703
0.17473 0.41876 0.91494
0.18919 0.41606 0.91282
0.2026 0.41335 0.91066
0.21513 0.41062 0.90848
0.22694 0.40787 0.90626
0.23813 0.40511 0.90401
0.24879 0.40233 0.90174
0.25897 0.39954 0.89943
0.26874 0.39672 0.89709
0.27814 0.39389 0.89472
0.28721 0.39105 0.89232
0.29597 0.38818 0.88989
0.30445 0.3853 0.88743
0.31268 0.3824 0.88494
0.32068 0.37948 0.88242
0.32845 0.37654 0.87987
0.33603 0.37359 0.87729
0.34341 0.37061 0.87468
0.35062 0.36761 0.87204
0.35766 0.3646 0.86937
0.36454 0.36156 0.86667
0.37128 0.35851 0.86394
0.37787 0.35543 0.86118
0.38433 0.35233 0.85839
0.39066 0.3492 0.85558
0.39687 0.34606 0.85273
0.40297 0.34289 0.84986
0.40895 0.3397 0.84695
0.41482 0.33649 0.84402
0.42059 0.33325 0.84106
0.42626 0.32998 0.83807
0.43184 0.32669 0.83505
0.43732 0.32338 0.832
0.44271 0.32004 0.82893
0.44802 0.31667 0.82582
0.45324 0.31327 0.82269
0.45838 0.30984 0.81953
0.46344 0.30638 0.81634
0.46842 0.3029 0.81313
0.47332 0.29938 0.80989
0.47816 0.29583 0.80662
0.48292 0.29224 0.80332
0.48761 0.28863 0.8
0.49223 0.28497 0.79665
0.49678 0.28128 0.79327
0.50127 0.27755 0.78987
0.50569 0.27379 0.78644
0.51005 0.26998 0.78299
0.51435 0.26613 0.77951
0.51859 0.26224 0.776
0.52276 0.2583 0.77247
0.52688 0.25432 0.76891
0.53094 0.25029 0.76532
0.53494 0.2462 0.76171
0.53888 0.24207 0.75808
0.54277 0.23788 0.75442
0.5466 0.23362 0.75074
0.55038 0.22931 0.74703
0.55411 0.22494 0.7433
0.55778 0.22049 0.73955
0.5614 0.21598 0.73577
0.56496 0.21139 0.73196
0.56848 0.20671 0.72814
0.57194 0.20196 0.72429
0.57536 0.19711 0.72042
0.57872 0.19217 0.71652
0.58204 0.18712 0.7126
0.58531 0.18196 0.70866
0.58852 0.17668 0.7047
0.59169 0.17127 0.70072
0.59481 0.16571 0.69671
0.59789 0.16001 0.69269
0.60091 0.15413 0.68864
0.60389 0.14806 0.68457
0.60683 0.14178 0.68048
0.61094 0.13685 0.67755
0.61625 0.13336 0.67579
0.62151 0.12978 0.674
0.62673 0.1261 0.67218
0.63192 0.12231 0.67034
0.63707 0.11841 0.66848
0.64217 0.11438 0.66659
0.64725 0.11021 0.66468
0.65228 0.1059 0.66275
0.65727 0.10142 0.66079
0.66223 0.09677 0.65881
0.66716 0.091915 0.6568
0.67204 0.086836 0.65478
0.6769 0.081501 0.65273
0.68171 0.075874 0.65065
0.68649 0.069904 0.64856
0.69124 0.063527 0.64644
0.69595 0.056656 0.64431
0.70062 0.049167 0.64215
0.70526 0.040873 0.63996
0.70987 0.031915 0.63776
0.71444 0.022992 0.63554
0.71898 0.014116 0.63329
0.72348 0.0052897 0.63102
0.72795 0 0.62874
0.73239 0 0.62643
0.73679 0 0.6241
0.74116 0 0.62175
0.74549 0 0.61939
0.7498 0 0.617
0.75407 0 0.61459
0.7583 0 0.61217
0.7625 0 0.60972
0.76667 0 0.60726
0.77081 0 0.60478
0.77492 0 0.60227
0.77899 0 0.59976
0.78303 0 0.59722
0.78703 0 0.59466
0.79101 0 0.59209
0.79495 0 0.5895
0.79885 0 0.58689
0.80273 0 0.58427
0.80657 0 0.58163
0.81038 0 0.57897
0.81416 0 0.57629
0.81791 0 0.5736
0.82162 0 0.57089
0.8253 0 0.56817
0.82895 0 0.56543
0.83257 0 0.56268
0.83615 0 0.55991
0.83971 0 0.55713
0.84323 0 0.55433
0.84671 0 0.55151
0.85017 0 0.54868
0.85359 0 0.54584
0.85698 0 0.54299
0.86034 0 0.54012
0.86367 0 0.53723
0.86696 0 0.53433
0.87022 0 0.53142
0.87345 0 0.5285
0.87665 0 0.52556
0.87982 0 0.52261
0.88295 0 0.51965
0.88605 0 0.51668
0.88912 0 0.51369
0.89215 0 0.5107
0.89516 0 0.50769
0.89813 0 0.50467
0.90107 0 0.50163
0.90397 0 0.49859
0.90685 0 0.49554
0.90969 0 0.49247
0.9125 0 0.4894
0.91528 0 0.48631
0.91802 0 0.48321
0.92074 0 0.48011
0.92342 0 0.47699
0.92606 0 0.47386
0.92868 0 0.47073
0.93126 0 0.46758
0.93381 0 0.46443
0.93633 0 0.46127
0.93882 0 0.45809
0.94127 0 0.45491
0.94369 0 0.45172
0.94608 0 0.44852
0.94844 0 0.44532
0.95076 0 0.4421
0.95306 0 0.43888
0.95531 0 0.43565
0.95754 0 0.43241
0.95974 0 0.42916
0.9619 0 0.4259
0.96403 0 0.42264
0.96613 0 0.41937
0.96819 0 0.41609
0.97022 0 0.41281
0.97222 0 0.40952
0.97419 0 0.40622
0.97613 0 0.40291
0.97803 0 0.3996
0.9799 0 0.39628
0.98174 0 0.39295
0.98354 0 0.38962
0.98532 0 0.38627
0.98706 0 0.38293
0.98877 0 0.37957
0.99044 0 0.37621
0.99209 0 0.37284
0.9937 0 0.36947
0.99528 0 0.36608
0.99682 0 0.36269
0.99834 0 0.3593
0.99982 0 0.35589
1 0 0.35248
1 0 0.34907
1 0 0.34564
1 0 0.34221
1 0 0.33877
1 0 0.33532
1 0 0.33187
1 0 0.32841
1 0 0.32493
1 0 0.32146
1 0 0.31797
I was really happy to discover just now that the shg function now works as it should with the new desktop implementation in R2025a. It is a quick way, using only the keyboard, to bring the current figure window forward. In the beta desktop in recent releases, it had no effect.
This function originated in 1980s MATLAB running on DOS, when there were no windows, and PC monitors had separate text and graphics modes that could not be viewed simultaneously. The function shg ("show graphics") would switch the monitor into graphics mode.
The function was retained after the text mode / graphics mode distinction became obsolete, and its meaning shifted to "bring the current figure forward."
It is 37-year-old keyboard muscle memory for me. 🙂
It looks like the default figure size has been increased.
R2025a:
>> get(groot,"DefaultFigurePosition")
ans =
308 245 1111 669
R2024a:
>> get(groot,"DefaultFigurePosition")
ans =
584 495 560 420
But more of the figure is going towards decorations / toolbars / menus now.
I believe those decorations / toolbars / menu are outside (above) the figure position rectangle. The portion of the overall window devoted just to the traditional figure and its contents is significantly bigger in R2025a. See the overlaid pixel ruler in this screenshot.
I just discovered that the figure toolbar is collapsible, leaving only the tabs showing. I'm not sure if this toolbar-collapsed state is remembered for future figures, or if it is programmatically settable via the settings mechanism.
I am happy to see the larger default figure size. I still do think that the margins are too large, though. I find myself manually reducing figure margin size A LOT.
@Steve Eddins: How did you overlay those pixel rulers?
@goc3, I use a Mac application called Free Ruler, which is available in the App Store. That and Digital Color Meter are my go-to tools for close inspection of what's on the Mac screen.
I just installed Free Ruler—it looks like it will be helpful.
The built-in Digital Color Meter is extremely useful. I have easily used it at least a hundred times.
@goc3, I agree about the margins. I just checked the default axes position, and it did not change. Since the default axes units are normalized, the bigger figure figure size (in pixel units) means that those margins are now bigger (in pixel units) in R2025a.
I was happy to discover that the problem with figure titles and axis labels being too small on the Mac has finally been addressed. It's a subtle but meaningful fix.
R2025a screen shot:
R2024a screen shot (after resizing the figure to be the same size as the new default in R2025a):
I'm really concerned about what happened to the Simulink Diff tool in this release:
Not only is text in embedded Matlab functions, Stateflow etc. truncated so that you can't even see the entire text or differences for that matter, you also can not select the text:
The text diffing of simulink and stateflow blocks was already abyssimal in the previous versions (need to set the coloumn count) but this is something else.
Please introduce scroll bars for text in simulink blocks! Otherwise this is not usable as a diff tool anymore!
Thank you very much.
I really hope that this gets fixed!
Any major perofrmance imporvements?
Maybe better memory handling in generated code?
A suggestion to improve the live editor:
Please allow us to have consecutive code but in different cells, currently one cannot have two consecutive code cells, they have to add a text cell in between. It will be helpful if you include this in next release.
Thank you.
I think you should be able to achieve this by adding a section break between the code sections:
I think I know what you mean. Consecutive code cells are automatically merged.
code cell number 1
code cell number 2
and you need text between in order to separate them
code cell number 3
code cell number 4
I attempted to reproduce this behavior, but I haven't seen it yet. My setup: 2021 MacBook Pro M1, R2025a Prerelease Update 1, live script as .mlx and live script as .m.
The issue is that you need to use code section then a text gap and then a code section. You cannot use code section and then immediately a different code section: the two code sections will automatically get merged together.
I am not sure why exactly you would want two adjacent code sections, but I acknowledge that they cannot currently be created.
I'm not seeing the merging of adjacent code sections in my setup (2021 MacBook Pro M1, R2025a Prerelease Update 1). Where are you seeing this happen?
Insert a line of code in the LiveScript editor using the "Code" ribbon tool.
Press the Text button to end the code box.
Now, click the "Code" button again while positioned on the text line, in hopes of starting a new code box.
The result is that the new code box is merged with the previous code box. There is no dividing line after the first line, no separation between the areas. In the MATLAB Answers editor, there is a single Copy button for the merged area, rather than seperate copy buttons.
You can sort of do what is desired by inserting a Section Break.
You can press new line and then start a code block, and it does not merge? Unless they fixed it in the meantime. Otherwise pressing text innthe middle of a code block also does this. Not sure what thr default behavious without new line should be
Pressing newline between is an example of "have to add a text cell in between". The request is for the ability to have two distinct consecutive code areas with no text area between them.
Personally I do not feel as if putting a blank line between them is a significant problem, but I am not the original poster.
Fortunately, R2025a has begun to incorporate full support for dark mode, but unfortunately, there is no built-in AI programming assistant like GitHub Copilot(Although the intelligent suggestions have been enhanced, they still fall far short of GitHub Copilot.).
I made it clear a few weeks ago that there would not be any copilot in r2025a.
(I do not know if or when such a feature will be introduced.)
👀
Stay tuned to the Graphics and App Building blog for R2025a coverage following the general release. I'm looking forward to reading more in this thread.
JavaFrame property, javacomponent function, actxcontrol function have all been removed.
This will be major for people like @Yair Altman
MATLAB no longer uses OpenGL® technology to render graphics
Very interesting! Any follow-up docs about that change, the implications, and the reasons behind?
Thanks :)
I get the impression that the underlying plotting technology behind figures has changed. I think this needs to be read in conjunction with javacomponent now having gone away -- an indication that now plotting no longer uses java (or OpenGL)
Does that have some effects on plotting objects with transparency (alpha<1)?
No, transparency levels and gradients should not be affected.
Sparse Matrix: Create and use single-precision sparse matrices could be very useful for people who use sparse arrays. (I do not happen to do much sparse array work myself.)
Reading Compressed Data: Read data from compressed and archived files will be useful sometimes.
Figure Container: Organize, explore, customize, and generate code for plotted data using an improved figure window
By default, figures are all docked inside a figure container (this is not the same thing as being docked to the command window.)
I don't especially like this docking, and I don't especially like the new big toolbar for figures. I am going to have to spend some time figuring out how to get around this.
Love that!
Unfortunately, performance is still a bit lower than using the old figures, especially when a lot of calculations need to be done after rescaling or axes limits changes. Happens a lot when using many axes elements like a plotmatrix.
I assume there's a lot of work going on to improve the snappiness?
However, apps and saved figures that modify the figure menu bar or toolbar (for example, by setting the ToolBar property of the Figure object) continue to display the menu bar and toolbar instead of the toolstrip. If your app relies on menu bar or toolbar functionality, you can specify that the app display the menu bar or toolbar instead of the toolstrip by setting the figure MenuBar or ToolBar properties when you create the figure for your app.
f = figure;
f.MenuBar = "figure";
f.ToolBar = "figure";
The maximum length of MATLAB identifiers increased to 2048 will be important to some people. (Not to me though.)
The following functions may come in handy on this end
This is very important for us (FSDA toolbox). When you receive data from official institutions the name of the variables exceeds the 63-character previous limit
I have been looking forward to this. I have run into the 63-character limit many times when naming test-class functions. It will be nice to be able to write more descriptive test-method names.
The new matlab.lang.Workspace ability to manipulate workspaces is intriguing.
Sounds like a great step towards more modularization. I'm highly curios what MathWorks' examples will look like.
I was around for design discussions about this and advocated for it. I’m excited to see that it has shipped.
It is on the list of blog articles to be published once R2025a has shipped.
cell2mat function accepts larger set of data types will be useful at times.
Finally, always used my own version of cell2mat for that purpose. Love that change.
One of the changes is the built-in support for Dark Mode and Themes.
Personally I find Dark Mode boring and hard on my eyes, but I know many people have been asking for it.
In the past, I have found that editing light text on a dark background was hard on my eyes, but I might try this again.
Select a Web Site
Choose a web site to get translated content where available and see local events and offers. Based on your location, we recommend that you select: United States.
You can also select a web site from the following list
Americas
- América Latina (Español)
- Canada (English)
- United States (English)
Europe
- Belgium (English)
- Denmark (English)
- Deutschland (Deutsch)
- España (Español)
- Finland (English)
- France (Français)
- Ireland (English)
- Italia (Italiano)
- Luxembourg (English)
- Netherlands (English)
- Norway (English)
- Österreich (Deutsch)
- Portugal (English)
- Sweden (English)
- Switzerland
- United Kingdom(English)
Asia Pacific
- Australia (English)
- India (English)
- New Zealand (English)
- 中国
- 日本Japanese (日本語)
- 한국Korean (한국어)