Topaz Video AI 5.1 - 5.1.2

Hello everyone!

Today we have a new release, Video AI 5.1.2


Changelog from 5.1.1

  • Fixed exports not appearing.
  • Fixed image sequence related crashing.


Changelog from 5.1

  • Fixed slow performance of exports.
  • Fixed Low Power Mode missing in preferences.


There are several major changes that we’re excited to announce:

1. Instant Rendering (Experimental)

Video AI now renders previews as soon as you click ‘Play’.

Instant renders make the timeline more interactive and allow you to directly compare models with shorter wait times.

2. Frame Interpolation for DaVinci Resolve OFX (macOS + Windows)

Access Apollo, Aion, and Chronos directly from DaVinci Resolve. Convert to slow motion at up to 16x interpolation and make use of Resolve’s retiming controls for smooth speed-ramping.

To use Frame Interpolation in Resolve, be sure to follow the guide posted here.

3. New multi-GPU options (Video AI Pro)

Video AI now includes two multi-GPU modes. The existing “All GPUs” option now supports NVIDIA cards running TensorRT models, leading to major utilization increases for systems with 2+ NVIDIA cards. This mode is now named “Single video” under “GPU Settings”.

The second mode, “Multiple videos”, optimizes for large export queues when running on systems with multiple GPUs. We’ll have more to share about this Video AI Pro feature when it launches in June.

4. New Welcome Screen

Quickly reopen projects and resume renders using the new Welcome Page. This is a home for recent projects and includes the ability to select favorites for easy access.

5. Preferences UI Refresh

Navigate settings using the new preferences sidebar with expanded categories and new tooltips.

6. Colorspace settings added to Video Input Options

Set custom colorspace, color primaries, color trc, and color range for video inputs with metadata issues.



@tony.topazlabs @kyle.topazlabs

Still has Greyed out options
you show.

1 Like

If you bothered to read the release notes you’d see this feature is not yet available and won’t be unless you buy a ‘pro’ licence once they start selling them…



If all the betas are public now, what’s the point of posting “release” versions with features that are still experimental and not ready for prime time?


why is this update making the app run sooooooo slooooooow? please tell me this gets resolved. what used to take 2 minutes is now taking 15. i’ve got a new mac studio, so it’s not hardware. wish i hadn’t updated.

1 Like

a 5 second preview is taking 8 minutes. i’m scared to actually export anything. this has brought my project to a halt. DO NOT DOWNLOAD

1 Like


Works fine for me, you probably have something misconfigured.


#1 priority should be updating models upscaling videos is the main reason 90% of people are using this program. Most users are not using slowmo or any of the addon plugins Video AI only get model updates 3 or 4 times a year. This is weak update could have waited 2 to 3 more weeks before dropping this version. I not trying to be rude I am saying what a lot of people are complaing about outside this forum.


the only thing i changed was getting this new version. just trying to do some denoising. something this program did in minutes before this version. yeah, my bad.

0.3 fps.


i can see that you’re trying to help me while i’m complaining and I appreciate it. i’ll look closer.

sadly, my prefs are looking normal, i’ve restarted and i’ve got an hour to go on outputting a 1:30 video file.

half way through!

Without knowing your previous performance levels and settings I can’t say if there’s an issue or not.

I’d need to see a benchmark result, and see screenshots of exactly what settings you are using for the export/previews.

If you changed anything since the last version you used (type of enhancement/model used, output settings, preferences settings) it’s likely you won’t get the same performance.

You’ll need to double check everything.

1 Like


1 Like

I would recommend you re-select your gpu in the processing prefs, and not rely on on “auto” if that was the setting.


I thought SLI GPU support was discontinued many years ago? If we were to use this feature with say, a 40 series cards, how do we connect two together or just connect them into motherboard the same way if you had a single GPU?

Thanks for the update… but now i’m waiting the PREMIERE PRO Plug-in! :sweat_smile:

1 Like

This seems to be a bug on the Mac version: in the new settings panel, there is no longer a toggle for ‘low power mode’. For me this means that topaz is now permanently stuck in processing in low power mode with no ability to toggle this off. Is there any fix for this currently other than downgrading to a previous version?

that doesn’t seem to be an option for me. in those prefs, the only options i have are max processes (set to 1) and max memory (100%). this is upending a handful of projects for me. is there a way to get back to the earlier version?

SLI is indeed not supported
Multiple GPU are recognised through the PCIE bus and Drivers (latest)
Moreover SLI is clearly recomended to give up on Nvidia site, this is for all Ampere >=

→ Also to coders of Topaz Ai Video are you aiming to update your models with space environement ?
No one of existign one doeas a correct job, or at most doeis it pushing details to over values that lead to introduce noise or artifact.

Just a warn than your direct concurrents did recently a huge step in this way. Topaz is at least 50% less efficient for 2x or 4x the price. And I don’t talk about tensor use, which is to my mind very far to be optimized. Jobs are still understood as 3D task and not computed ones.
CUDA 12.4 is now a must have as older one are not anymore supported.

GPU load does not reach 100%, this it not a normal behaviour on a perfect use of CUDA.

  • CPU Load : Why Topaz doesn’t use with better efficiency tasks given to CPU
    Threaripper 2990WX is almost sleeping with 2K to 4K upsclaing, concucrents use it almost @ 75%

  • GPU Load : At most 75 to 98% in all cases/situations when using 1 to 3 of them.
    Same happen on Nvidia or AMD, but Nvidia performs far better

  • RAM use : Why so few memory use ? Workstation have quad channel with huge amount of it. This could at least be configurable. Setting 64 GB state of the art RAM lead to the same than 32 GB premium RAM.

  • Dedicated cache SSD for temporaray jobs : fully sleeping, used at 3% (7Gbps write sequential, 70k IO)

I even can’t imagine users that are using gaming settings whoch sould be most of the cases. → app should be unusable :confused:
I reach at most 30 fps of rendering with some high end quality settings. Some of your concurents reach 60 fps on a single GPU :S

Best regards