NVIDIA's latest driver quietly adds the first signs of DLSS 5 Neural Rendering support

NVIDIA has released GeForce Game Ready Driver 610.47, and buried inside it are the first signs of DLSS 5 making its way into the driver pipeline. User Warkratos discovered through the NVIDIA Profile Inspector that the new driver appears to add three new profile entries: DLSS-NR, DLSS-NR Streamline, and DLSS-NR Presets. The "NR" part almost certainly stands for Neural Rendering, which is how NVIDIA has been presenting DLSS 5 since its GTC 2026 reveal.

Before anyone gets too excited, these entries do not make DLSS 5 usable in any current games. Even if you enable the preset through the NVIDIA Profile Inspector, you will not get any DLSS 5 effects in existing titles. NVIDIA has not yet shipped any DLSS 5 files to go alongside these flags, so for most users, this is early groundwork rather than anything immediately useful.

3

For context, DLSS 5 introduces a real-time neural rendering model that takes a game's rendered 2D frame and motion vectors as input, then uses an AI model to infuse the scene with photoreal lighting and materials anchored to the source 3D content. NVIDIA's pitch is that it delivers a level of photorealistic computer graphics previously achievable only in Hollywood visual effects, running in real time at up to 4K resolution.

The technology drew significant backlash following its early demos, with many gamers criticizing the altered visuals and dubbing them "AI slop." But despite the harsh criticism, NVIDIA isn't backing off, and its plans to win over the more than 50% of users who don't want DLSS 5 altering their games involve assuring developers they will retain full artistic control.

3

The new driver updates suggest the 600-series driver will add more complete DLSS 5 support later this year. NVIDIA has confirmed a Fall 2026 launch window for DLSS 5, with titles including Phantom Blade Zero, Resident Evil Requiem, Assassin's Creed Shadows, Hogwarts Legacy, Starfield, and The Elder Scrolls IV: Oblivion Remastered among the first to support it. Hardware requirements are not yet confirmed, but the technology will likely only work with RTX 50-series cards, since early demos used two RTX 5090 cards, and NVIDIA is trying to reduce that to just one.