Apex Neural Rendering Could Be a Serious NVIDIA DLSS Alternative for Frame Generation
NVIDIA DLSS may be facing a new challenger in the neural rendering space. Apex Compute has introduced Apex Neural Rendering, or ANR, a cross-platform rendering suite designed to improve frame generation, image quality, and GPU memory efficiency across a wide range of graphics hardware.
The most exciting part is that gamers do not have to wait for the full package to arrive. ANR’s Frame Generation feature is already available in beta, giving users an early chance to test the technology in supported games.
Apex Neural Rendering is built around three major features: Frame Generation, Super Resolution, and Ray Reconstruction. Frame Generation is the first component available to the public, while the upscaling and ray reconstruction models are still in development. If Apex Compute can deliver on its claims, ANR could become one of the most interesting NVIDIA DLSS alternatives for gamers using both new and older GPUs.
One of the biggest advantages of ANR is its broad hardware support. Unlike NVIDIA DLSS Frame Generation, which is officially limited to newer RTX graphics cards, ANR is designed to work across a much wider selection of GPUs.
Supported NVIDIA hardware includes GeForce GTX 16 Series cards as well as RTX 20, RTX 30, RTX 40, and RTX 50 Series GPUs. AMD support covers Radeon RX 5000, RX 6000, RX 7000, and RX 9000 Series graphics cards, along with the Steam Deck. Intel users are also included, with support for Arc A-Series and B-Series GPUs.
That wide compatibility could make ANR especially appealing to PC gamers who want frame generation features but do not own the latest NVIDIA graphics cards. It also gives AMD, Intel, and handheld gaming users another option in a market that has often been dominated by proprietary GPU features.
Platform support is also promising. The ANR Frame Generation beta works on Windows 10 and Windows 11, while Linux and SteamOS are supported through Proton. Apple silicon is supported separately through a native Metal test kit for macOS 26 and newer.
Performance is one of the main selling points. According to Apex Compute’s internal benchmarks, ANR Frame Generation can be faster than NVIDIA DLSS Frame Generation in certain scenarios. On a GeForce RTX 5090 running at 3840×2160 resolution, ANR takes 1.01 milliseconds per generated frame in 2X mode, compared with 1.26 milliseconds for NVIDIA DLSS Frame Generation.
In 4X mode, Apex reports that ANR requires 0.61 milliseconds per generated frame, while NVIDIA’s solution takes 0.82 milliseconds. Apex Compute also claims that ANR can deliver up to 2.4 times faster frame generation in some tests.
However, it is important to understand what these numbers mean. These results refer to the processing cost of generating additional frames, not a guaranteed 2.4 times increase in actual in-game frame rates. Real-world performance will depend on the game, GPU, CPU, settings, resolution, and overall system configuration.
GPU memory usage is another area where ANR appears to have an advantage. Apex Compute says ANR uses just 214 MB of GPU memory at 4K, compared with 562 MB for NVIDIA’s frame generation technology. That works out to roughly 62% lower memory usage. The neural network weights are also extremely small, reportedly taking up only 0.17 MB.
Lower VRAM usage could be important for gamers using mid-range or older graphics cards, especially as modern PC games continue to demand more memory at higher resolutions and texture settings. If ANR can maintain good image quality while using significantly less GPU memory, it may be useful for systems where every megabyte matters.
Image quality is another major focus for Apex Compute. The company has shown side-by-side comparisons designed to highlight difficult frame generation scenarios, including fast-moving objects, particles, water, vegetation, animated characters, and disocclusion artifacts. These are the kinds of visual challenges where frame generation systems can sometimes create ghosting, distortion, flickering, or other unwanted artifacts.
Apex Compute claims ANR performs better than NVIDIA DLSS Frame Generation in internal quality tests using metrics such as CGVQM-5, ColorVideoVDP, and PSNR. These comparisons reportedly favor ANR across 2X to 6X multi-frame generation modes.
The company has also shown ANR running in games such as The Last of Us Part II Remastered and Hogwarts Legacy. While the early demonstrations look promising, independent testing will be essential before drawing firm conclusions. Internal benchmarks can be useful, but real-world analysis from a wider variety of games and hardware configurations will determine how strong ANR really is.
Gamers interested in testing ANR Frame Generation can download the latest beta from the project’s GitHub repository. Installation uses OptiScaler, with separate packages depending on whether a game already has OptiScaler installed or not. The basic process involves extracting the files, running the installer against the game’s executable directory, and then enabling multi-frame generation through the game’s graphics settings.
The latest Frame Generation beta release, version B14, was published on November 12. This update fixes a Windows compatibility problem that caused some games to fail to launch when older Visual C++ runtimes were present.
Apex Compute has also released a utility called nr-record. This tool lets users capture frame generation inputs and replay them through competing implementations, making it easier to compare image quality directly. For enthusiasts and developers, that could be a valuable way to evaluate artifacts, motion stability, and visual consistency across different frame generation systems.
There are still some important limitations. ANR’s multi-frame generation implementation is not currently open-source, although its developers say the source code will be released later. Since the current version is still an experimental beta, users should also expect bugs, compatibility problems, and possible visual artifacts in some games.
Even with those caveats, Apex Neural Rendering is an exciting development for PC gaming. A frame generation system that works across NVIDIA, AMD, Intel, Steam Deck, Linux, SteamOS, Windows, and Apple silicon could be a major step toward more open and accessible neural rendering technology.
For gamers with older graphics cards, ANR may be especially interesting. Many users have been looking for ways to bring advanced frame generation features to hardware that does not officially support the latest GPU technologies. If Apex Compute continues improving performance, compatibility, and image quality, ANR could become a serious alternative in the growing world of AI-powered gaming graphics.
