Tuesday, July 21st 2026
NVIDIA Shows DLSS 5 Progress and Technical Details at SIGGRAPH 2026
NVIDIA has once again showcased its DLSS 5 neural rendering at the SIGGRAPH 2026 event, currently taking place in Los Angeles. Despite the initial negative reaction from gamers, NVIDIA is continuing to develop the technology, offering improvements and a new demonstration of how DLSS 5 is intended to function. One of the most interesting aspects is that game developers can choose from three DLSS 5 models, each providing a different level of structural intensity, global illumination, texture detail, and more. These models, labeled Model A, Model B, and Model C, can be tailored to specific scenes or character needs, allowing developers to mix and match according to their requirements. Each model carries a different parameter count, meaning that some are more compute and memory-intensive than others. However, they generally fit and run fine on a single GPU without any problems.
Another noteworthy feature is the preservation of artistic intent, which was heavily criticized in the original DLSS 5 demo. Each individual object, scene, and game setting can be customized according to the developer's vision. This is one of the challenges NVIDIA faced with DLSS 5, as maintaining visual identity, object semantics, and character poses, along with output lighting semantics, motion, and art direction, is crucial. Within these strict parameters, DLSS 5 applies material photorealism to enhance visual appeal. Below you can even see the customization of the applied model, where there are adjustable parameters that can be done per scene and per character.The second challenge that the company faced is temporal streaming. Traditional generative AI video generators process in chunks, reasoning across a frame and generating a batch of frames at once. However, real-time rendering requires immediate generation, frame by frame. Since DLSS 5 uses motion vectors from the game engine, it can generate additional textures instantly, without delay. This results in no shimmer or image corruption, as DLSS 5 know exactly where to attach itself and perform its texture generation instantly.The final challenge for DLSS 5 is speed. To render a 4K frame with 8.3 million pixels at over 60 FPS, NVIDIA achieves this in less than 16 ms, with most of the processing time consumed by the game engine rather than the DLSS 5 model. The model is so efficient that users will not notice any delay during gameplay, and it operates on a single GPU. While NVIDIA highlights VRAM efficiency, it remains unclear whether 4K gaming will be limited to high VRAM GPUs or something else. As the launch is expected this fall, the company will provide general system requirements and further details, but the improvements compared to the original demo seem promising for now.
Source:
NVIDIA at SIGGRAPH 2026
Another noteworthy feature is the preservation of artistic intent, which was heavily criticized in the original DLSS 5 demo. Each individual object, scene, and game setting can be customized according to the developer's vision. This is one of the challenges NVIDIA faced with DLSS 5, as maintaining visual identity, object semantics, and character poses, along with output lighting semantics, motion, and art direction, is crucial. Within these strict parameters, DLSS 5 applies material photorealism to enhance visual appeal. Below you can even see the customization of the applied model, where there are adjustable parameters that can be done per scene and per character.The second challenge that the company faced is temporal streaming. Traditional generative AI video generators process in chunks, reasoning across a frame and generating a batch of frames at once. However, real-time rendering requires immediate generation, frame by frame. Since DLSS 5 uses motion vectors from the game engine, it can generate additional textures instantly, without delay. This results in no shimmer or image corruption, as DLSS 5 know exactly where to attach itself and perform its texture generation instantly.The final challenge for DLSS 5 is speed. To render a 4K frame with 8.3 million pixels at over 60 FPS, NVIDIA achieves this in less than 16 ms, with most of the processing time consumed by the game engine rather than the DLSS 5 model. The model is so efficient that users will not notice any delay during gameplay, and it operates on a single GPU. While NVIDIA highlights VRAM efficiency, it remains unclear whether 4K gaming will be limited to high VRAM GPUs or something else. As the launch is expected this fall, the company will provide general system requirements and further details, but the improvements compared to the original demo seem promising for now.











153 Comments on NVIDIA Shows DLSS 5 Progress and Technical Details at SIGGRAPH 2026
It's like saying that in Blender, Eevee + ReShade can replace Cycle. More things are going on with DLSS 5 than better contrast and sharpness
Can ReShade edit images like this
AFMF is based on frame interpolation rather than AI frame generation, so the quality of the interpolated frames is poor.
And just leaving it off if I don't like it. remember, That's always an option ;)
Control what you can control, you won't change the minds of people you perceive to be Nvidia fanboys acting like this, quite the opposite even.
I’ve always called Frame gen “trash” no matter what brand the trash bag is, it’s still trash.
A real breakthrough isn’t integrating image diffusion into upscaling, a breakthrough is having hardware that doesn’t need upscaling to deliver 4K, 60–120 fps.
A real breakthrough would be games that don’t use TAA, so the baseline would actually be native. But nah, now that we have a crutch, let’s stop investing in proprietary engines and simply use UE5 underutilizing the hardware. The experience of beta testers who buy the product doesn't matter.
It's actually kinda crazy what you can do with reshade."Frame interpolation" just means that the frame inserted between frames are based on frames before and after. "Frame Generation" is mostly just newer techniques/methods that make better interpolated frames. And that said... AFMF does use Frame Generation techniques, it just not as good as game side DLSS, FSR or XeSS Frame-Gen. Which makes sense, as when its game-side it can feed the methods allot better data and produce nicer looking interpolated frames so it can look smoother with less artifacting.
This simply enables bad optimization normalization.
Publishers now routinely ship games expecting upscaling and frame generation to save their poor performance. Devs don’t optimize the base engine code anymore.
They depend on DLSS, FSR, or XeSS to save their hides.
I don't think DLSS/MFG tech is trash, it gives life to old/slower hardware like laptops, consoles, and handhelds. It even allows non flagship cards to reach flagship frames. The latency is subjective its not noticeable in my experience but have not had the displeasure of playing any titles below 60fps this generation yet. Also even E-Sports titles can be run by smoothly by APUs nowadays without the 'magic'.
Honestly the less people are blindly following the internet's negative opinions the more will realize they don't need to upgrade thanks to these new techniques. It could've been worse, NVIDIA and AMD focus on Raster just to be able to run Unreal Engine 5 games at 50-60fps then all the previous generation hardware is stuck at 20-30fps .. actually that's happening now 😆
Bad optimisation isn't a new phenomenon either, nor is midrange or older hardware getting left behind, in fact that used to happen way faster and FPS overall was generally lower.
Game devs work with what they have, in fact one could argue games look better today than they would have if excellent upscaling options didn't exist.
Honestly, I feel like I am taking crazy pills sometimes.
Um, you should know that what Reshade can do is very limited, as it is completely post-processing, with effects far inferior to DLSS5. Reshade RTGI is improving the effects of old games, with not much it can do, let alone surpassing modern games. When old games were created, relevant technologies were not popular, which is why there was room for improvement, whereas now RT has become the graphics standard for DX12, making these restricted Reshade RTGI completely unnecessary.
DLSS5 its at post-processing, but does the work of the rendering pipeline, and traditionally achieving the effects of DLSS5 could not be processed in real-time at all, as AI deeply processes the lighting, material rendering, textures, and subsurface reflections, which is different from simple Reshade effects.With TAA, 1440p only offers 480p clarity, until DLSS appeared, allowing 1440p to reach 4k clarity, and the latest DLSS model can achieve near 4k using just 1080p. Then with the launch of MFG, the display refresh rate rather than the graphics card became the limit. FG added 5ms of latency, although the latency was completely offset by DLSS and Reflex and is far better than the latency provided by consoles, NVIDIA is just too evil.