Monday, June 22nd 2026

Leaked FSR 4.1 INT8 Model Confirms Compatibility with RDNA 3.5

Valve has begun integrating AMD's FSR Upscaler 4.1 into its Proton compatibility layer to enable FSR 4.1 on the SteamOS Linux-based operating system. Recently, we reported that AMD has been developing a separate model for FSR Upscaling 4.1 to function on RDNA 3. This involves converting from the FP8 data type used on RDNA 4 GPUs to INT8 data types to ensure smooth operation on RDNA 3. One of the biggest questions surrounding FSR 4.1 compatibility was whether the solution would work with RDNA 3.5 integrated graphics, in addition to the already announced RDNA 3 IP support. Valve has now leaked the FSR 4.1 INT8 model, which has been observed running on RDNA 3.5 thought Radeon 890M GPU.

Valve has included preliminary FSR 4.1 support in its Proton Experimental build, which is used for testing features before they appear in the stable release that ships with SteamOS builds. A Redditor managed to download the leaked file before Valve removed it and ran a few tests using OptiScaler to force different FSR DLL versions in games. Since the file is officially signed, games checking for DLL signatures reportedly ran without issues. This is great news for those waiting for the official DLL release, especially RDNA 3.5 users who were left uncertain about whether FSR 4.1 would work with RDNA 3.5.
Finally, we understand that there are some fundamental differences in the RDNA generations regarding microarchitecture. For RDNA 3, there is no support for 8-bit floating point, as only the latest RDNA 4 uses FP8. To run the FSR 4.1 Upscaling model on RDNA 3, AMD must convert it to 8-bit integer data, as RDNA 3 hardware uses INT8 data types without FP8 support. Generally, integer math is sufficient to optimize these models, but the conversion work is necessary, which explains the delay between RDNA 4 and RDNA 3 support. However, since RDNA 3.5 appears to be closer to pure RDNA 3 with its data types support than RDNA 4, running the INT8 model of FSR 4.1 without any problems seems to be a good sign that the official support will be there.
Source: Reddit
Add your own comment

27 Comments on Leaked FSR 4.1 INT8 Model Confirms Compatibility with RDNA 3.5

#1
sam_86314
First FSR4.0 INT8 was leaked, now FSR4.1 INT8 has leaked.

I even tried FSR4.0 INT8 on an Intel Arc GPU back when the Optiscaler team made it really easy to use, and while the performance was garbage and it had graphical issues, it did technically work.

It's almost like AMD should've just released this already, or at the very least, said they were working on it.
Posted on Reply
#2
Event Horizon
Valve integrating it into Proton is the strongest sign the Steam Machine is getting supported.
Posted on Reply
#3
Rulumi
I think the problem in Windows was more of AMD adding specific checks for checking integrated graphics and potentially block execution if detected. Hopefully they don't end up actually blocking RDNA 3/.5 integrated graphics at release in Windows.
Posted on Reply
#4
ToTTenTranz
What's really interesting about this release is how it's not using WMMA (dual-issue instructions for ML loads on RDNA3+), so it's functionally identical to the old FSR4.0 leak from 2025.
It should therefore work on all RDNA GPUs.
Posted on Reply
#5
Macro Device
ToTTenTranzon all RDNA GPUs.
Could potentially be false for RX 5000 series but my RDNA2 card runs it fine. Not worth the upgrade from previously leaked FSR4 it seems though.
Posted on Reply
#6
outlw
This is good news!

I may be dreaming... but I do wish Valve would work some magic, and integrate Optiscaler styled functionality directly into SteamOS.
Being able to spoof the GPU/IGP, intercept all DLSS/XeSS/FSR2+3 calls, and convert it into working FSR4.1 would be a massive win for them, and Linux gaming in general.
Posted on Reply
#7
obsidian24776
26.6.2 is out. FSR 4 on RDNA 3 is launched with this driver.
Posted on Reply
#8
_roman_
really amuses me all the time when i read such low quality news pieces when i coded on a 8 bit ucontroller which was only able to use integer.

bullshit / nonsense / fake claims by AMD as always. there are no technical limitations. it is just a limit from the compiler and the guy who is not willing to convert the code or find an algorithm

How can someone do floating maths? I really wonder. Do I really need float when i have 8bit integer? Stupid AMD
  • no hardware subtraction circuit: subtraction is simulated via addition logic.
  • no hardware multiplication: requires software rotation and addition loops.
  • no hardware division: requires software rotation and subtraction loops.
  • no hardware signed arithmetic: requires manual software evaluation of the status register flags.
  • no hardware floating point unit: limited strictly to 8 bit integer operations.
Maybe time for a amd meme, "stupid, stupider, AMD".
edit: someone else was faster than I with that quote
Posted on Reply
#9
Oberon
_roman_really amuses me all the time when i read such low quality news pieces when i coded on a 8 bit ucontroller which was only able to use integer.

bullshit / nonsense / fake claims by AMD as always. there are no technical limitations. it is just a limit from the compiler and the guy who is not willing to convert the code or find an algorithm
Spoken like the guy whose ONLY experience is coding on something like a ATMega. You have literally no idea what you're talking about.
Posted on Reply
#10
Jtuck9
outlw6669This is good news!

I may be dreaming... but I do wish Valve would work some magic, and integrate Optiscaler styled functionality directly into SteamOS.
Being able to spoof the GPU/IGP, intercept all DLSS/XeSS/FSR2+3 calls, and convert it into working FSR4.1 would be a massive win for them, and Linux gaming in general.
An "official" form of something like "Optiscaler Client" would be nice, outside of AMD / Windows integration, on those fronts.
Posted on Reply
#11
aktpu
_roman_fake claims by AMD as always
But AMD hasn't claimed anything, this was a leak?
Posted on Reply
#12
TheinsanegamerN
OberonSpoken like the guy whose ONLY experience is coding on something like a ATMega. You have literally no idea what you're talking about.
Keep in mind he also claims that AMD's drivers are so bad they fill his hard drive with error reports, and he has to keep deleting them....instead of fixing whatever is generating hundreds of gigabytes of error messages.
RulumiI think the problem in Windows was more of AMD adding specific checks for checking integrated graphics and potentially block execution if detected. Hopefully they don't end up actually blocking RDNA 3/.5 integrated graphics at release in Windows.
The real issue here is AMD's disastrous PR. If they had just come out originally and said they were working on a version of FSR4 for older cards, but since they didnt have the FP8 hardware it was going to take time, people would have no issue with this.

Instead, AMD chose to claim they were not supporting older hardware after code leaked that showed they could, and they continue to plant their boots firmly in their mouth at every opportunity. So now they have a ton of scrutiny from the pissed off community that was created entirely of their own doing.
Posted on Reply
#14
Rulumi
Since Valve already released it, there was probably little sense for AMD to keep the original schedule.

In any case and of note it's that now AMD specifically says APUs (RDNA 3 specifically) will be getting FSR 4.1 later instead of vagueness. I wonder if it indeed is blocked for now on Windows with integrated graphics or if they let users of them to just run the "unoptimized model".

Posted on Reply
#15
Tek-Check
_roman_Maybe time for a amd meme, "stupid, stupider, AMD".
Coffee too sour today? A heatwave kicking in?
Posted on Reply
#16
Macro Device
Tek-CheckCoffee too sour today? A heatwave kicking in?
Anyone with a little bit of sense is displeased with AMD to say the least, I'll grant Roman that.

I have a 6700 XT. A card that despite physical capability to run FSR4 in ever so slightly hindered mode and despite urgency for it (let's be real, this GPU barely keeps up with modern games at native resolutions if we don't count obsolete ones like 900p and 720p), there's still no official support for it. None. Nada. Zilch. My GPU is just five-something years old and it's already treated like it was made when King Arthur was still a thing.

All that while RTX 2000 cards, two years older cards, still have full and complete support for any DLSS upscaling the first day these new libraries are released. Why would I pick AMD for my next purchase? AMD clearly don't give me any meritable "why".
Posted on Reply
#17
Hecate91
Macro DeviceAll that while RTX 2000 cards, two years older cards, still have full and complete support for any DLSS upscaling the first day these new libraries are released. Why would I pick AMD for my next purchase? AMD clearly don't give me any meritable "why".
Except the latest DLSS will run poorly on RTX 2000 series cards, and VRAM will still be an issue on everything but a 2080Ti because upscaling isn't a magical fix for a lack of VRAM.
Also not to mention anything before 50 series won't get the latest frame gen features or whatever Nvidia decides to sell their users on next.
Why an AMD card? I personally value having better price/performance, a melt free power connector, and drivers that aren't vibe coded. I don't care much for upscaling, it's better late than never.
Posted on Reply
#18
Tek-Check
Macro DeviceWhy would I pick AMD for my next purchase? AMD clearly don't give me any meritable "why".
It's up to you, really. I had 6800XT in one of my systems and used it happily for 5 years. The main reason I did not buy 3080 was terrible VRAM offer. I sold it and replaced it with 9070XT. Got really good price for the old card on eBay and just added some money to buy a new one. This is what I would advise you to do if you can't wait for this specific feature. I always sell cards into secondary market after roughly two generation and replace them, to keep up with new architectures.
Macro DeviceI have a 6700 XT.
There must be growing number of games where 12GB card has aged better than, say 8GB 3070. Despite the fact that upscaling is currently better on 3070, I am sure that more games are starting to choke in 8GB VRAM buffer. I'd like to see some updated testing of those two cards, especially in more memory challenging games. Perhaps you don't play some of those, but others do. So, it's never that straight-forward, depending on case use.
Macro DeviceMy GPU is just five-something years old and it's already treated like it was made when King Arthur was still a thing.
It will certainly take a bit longer to deliver the upscaler to RDNA2 cards. We know two things. 1. It will be delivered. 2. More work is needed to get it delivered due to architectural differences. We know the reasons. RDNA3 support was supposed to come at some point in July. It was delivered today, and more work is being done in the background for APUs too. Perhaps my laptop 6800H APU also gets it next year. Nice addition, but I don't use the feature much.
Posted on Reply
#19
Macro Device
Hecate91Except the latest DLSS will run poorly on RTX 2000 series cards
Running poorly is better than not running at all.
Hecate91VRAM will still be an issue on everything but a 2080Ti because upscaling isn't a magical fix for a lack of VRAM.
In many cases with RTX 2000 cards, main bottleneck is compute, not VRAM. It's 3070/4060/3080/5060 cards that have too few gigs for what they can do.
Hecate91better price/performance
Unless you live in fantasy world where AMD cards sell for their MSRPs since day 1, I don't see this as a valid reason. It was margin of error in pure raster and usually slight NVIDIA wins in ray tracing for all I care. Also there's no bang per buck in LLMs or any other CUDA-oriented stuff if you're on AMD.
Hecate91drivers that aren't vibe coded
I've seen those for AMD too. Both parties are equally guilty.
Hecate91melt free power connector
The only valid point. This connector is worse than hell, I agree wholeheartedly. Wish it never existed.
Tek-CheckThis is what I would advise you to do if you can't wait for this specific feature.
Nah, all AMD GPUs underdeliver in what I do. My next purchase is an NVIDIA card when I stop being flat broke.
Tek-CheckThere must be growing number of games where 12GB card has aged better than, say 8GB 3070.
Of course but there's also an increasing number of games poorly optimised for older RDNA architectures. While I theoretically get more stable framerates due to no OOM, I also get sub par experience because RDNA2 was an afterthought so it's almost equally rubbish on both GPUs. Pick your poison, shall we say.
Tek-CheckMore work is needed to get it delivered due to architectural differences.
Even an unpolsihed leak of FSR4.0 is leagues ahead of any FSR3.1 or XeSS implementation. We could've literally had it A YEAR AGO with AMD working on a "perfect" version in the background while their customer base is happy with what's essentially a massive upgrade.
Posted on Reply
#20
Tek-Check
Macro DevicePick your poison, shall we say.
I did, consistently over last 6 years, and never regretted buying RDNA2-3-4 cards with enough VRAM for my needs.
Macro DeviceEven an unpolsihed leak of FSR4.0 is leagues ahead of any FSR3.1 or XeSS implementation.
You can install publicly available injection files and use it until official version lands. What's stopping you?
Macro DeviceWe could've literally had it A YEAR AGO with AMD working on a "perfect" version in the background while their customer base is happy with what's essentially a massive upgrade.
No, we could not have it earlier because there is Sony in the picture, who co-developed the upscaler with AMD for their PS5 Pro console. They most likely put IP restrictions on the roll out schedule, first the console, then everything else a few months later, which is exactly what happened. If you were Sony, you would never give a free access to the upscaler to Microsoft, the company that did not invest one cent into development effort. Sony is the key to understanding of the timescale of the upscaler roll out.
Posted on Reply
#21
Hecate91
Macro DeviceRunning poorly is better than not running at all.
I don't see much point in having the newest upscaling if it it's unusable.
Macro DeviceIn many cases with RTX 2000 cards, main bottleneck is compute, not VRAM. It's 3070/4060/3080/5060 cards that have too few gigs for what they can do.
Most are going to have 2070,3070, or 4070, which all could've used more VRAM. Some of the lower tier cards happen to be compute or bandwidth limited for product segmentation.
Macro DeviceUnless you live in fantasy world where AMD cards sell for their MSRPs since day 1, I don't see this as a valid reason. It was margin of error in pure raster and usually slight NVIDIA wins in ray tracing for all I care. Also there's no bang per buck in LLMs or any other CUDA-oriented stuff if you're on AMD.
I don't buy any hardware in day 1, the 9070XT is approximately half the price of a 5070Ti, and I personally have no need for LLM's or CUDA, to each their own though.
Nothing stops you from running an LLM on a Radeon card, Nvidia has people locked in with CUDA though, and IMO they shouldn't be allowed that monopoly with how much control over the market Nvidia has at this point.
Macro DeviceI've seen those for AMD too. Both parties are equally guilty.
AMD has their slip ups, but Nvidia has been having worse driver quality with the 50 series, bad mistakes that shouldn't happen like crashing or fans not working.
Macro DeviceThe only valid point. This connector is worse than hell, I agree wholeheartedly. Wish it never existed.
Yeah, the connector is one reason why I refuse to give Nvidia any money.
If EVGA still made graphics cards at least there would've been a decent warranty policy in place, but I shouldn't have to rely on a warranty, or what board partners are doing with band aid fixes to the cable.
Macro DeviceEven an unpolsihed leak of FSR4.0 is leagues ahead of any FSR3.1 or XeSS implementation. We could've literally had it A YEAR AGO with AMD working on a "perfect" version in the background while their customer base is happy with what's essentially a massive upgrade.
You could've been using the "leaked" FSR4 a year ago with Optiscaler, for whatever reason it took AMD time to get the INT8 conversion right, and now official FSR4 for RDNA3 is better and faster than the leaked community version.
Posted on Reply
#22
TheinsanegamerN
Hecate91Except the latest DLSS will run poorly on RTX 2000 series cards,
Whether its performance is acceptable is up to the consumer to decide. Unlike AMD, nVidia gave us the choice of using it. It is optional, BTW. You do not have to use it if you dont want to.
Hecate91and VRAM will still be an issue on everything but a 2080Ti because upscaling isn't a magical fix for a lack of VRAM.
It works fine on 8GB 4060s, so it will also work on 8GB 2070s.
Hecate91Also not to mention anything before 50 series won't get the latest frame gen features or whatever Nvidia decides to sell their users on next.
Because hardware has advanced. Why do you guys keep throwing out this red herring argument thinking it makes you look better?

FSR4 INT8 works on older cards. FSR4 frame gen doesnt, and nobody expects FSR4 frame gen on rDNA 3 or 2. Please stop being disingenuous.
Hecate91Why an AMD card? I personally value having better price/performance, a melt free power connector, and drivers that aren't vibe coded. I don't care much for upscaling, it's better late than never.
Cool. Glad you're OK with worse support. Most consumers are not, which would inform AMD's abysmal market share compared to nVidias.

BTW, if you think AMD isnt using AI coding on their drivers, I have a bridge to sell you. They talk about AI more then nVidia do.
Posted on Reply
#23
JustBenching
Macro DeviceAnyone with a little bit of sense is displeased with AMD to say the least, I'll grant Roman that.

I have a 6700 XT. A card that despite physical capability to run FSR4 in ever so slightly hindered mode and despite urgency for it (let's be real, this GPU barely keeps up with modern games at native resolutions if we don't count obsolete ones like 900p and 720p), there's still no official support for it. None. Nada. Zilch. My GPU is just five-something years old and it's already treated like it was made when King Arthur was still a thing.

All that while RTX 2000 cards, two years older cards, still have full and complete support for any DLSS upscaling the first day these new libraries are released. Why would I pick AMD for my next purchase? AMD clearly don't give me any meritable "why".
I have a much more recent RDNA 2 card (2022 release date) and I got 0, zilch, nada, so don't complain too much :D

TBF, as others have pointed out before, we kinda knew it when we bought an amd gpu, so it's on us. At least that's the narrative around here, that amd is so bad at supporting it's own products that we should have known. They are not entirely wrong.
Posted on Reply
#24
chrcoluk
Chances of steam deck getting FSR4?
Posted on Reply
#25
outlw
chrcolukChances of steam deck getting FSR4?
Officially?
No idea, though I hope it comes (with the option to use the less demanding FSR 3.1).

Community mod driven?
Already there!
Posted on Reply
Add your own comment
Nov 18th, 2026 14:19 PST change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts