NVIDIA RTX Spark Gets DLSS 5, On-Device AI Characters and Neural Rendering Tools

What RTX Spark is

According to NVIDIA, RTX Spark is a Windows PC platform for AI development, creation and gaming. It is powered by an NVIDIA Blackwell RTX GPU and an NVIDIA Grace CPU, and it combines up to 128 GB of unified memory with the full NVIDIA CUDA stack and RTX technologies including ray tracing and DLSS 5.

The company says RTX Spark benefits from the full suite of DLSS technologies in supported games: Super Resolution, Ray Reconstruction, Dynamic Multi Frame Generation and 3D-Guided Neural Rendering. Dynamic Multi Frame Generation can generate up to 5 additional frames per rendered frame and adjusts the multiplier automatically. The latest NVIDIA app beta also adds a faster Performance Mode (Preset C) for Frame Generation, with minor trade-offs in image quality.

DLSS 5 and the Unreal Engine plugin

DLSS 5 with 3D-Guided Neural Rendering is supported on RTX Spark, starting with NBA 2K27. NVIDIA says an October model update runs 2x faster than the September version and halves GPU overhead while keeping comparable visual fidelity. It also reduces VRAM use at 4K by 20%. The model will ship by default in future DLSS 5 games.

For developers, DLSS 5 adds a final stage to the graphics pipeline that enhances lighting and material detail. The engine-rendered frame remains the foundation, and developer controls decide what can change. NVIDIA says Avalanche Software used the SDK controls to tune its integration to keep the artistic intent and style of Hogwarts Legacy.

The updated DLSS 4.5 plugin for Unreal Engine adds, in version 8.9.0, official Windows on Arm support for Unreal Engine 5.8. Native ARM64 binaries and support for building and packaging UE 5.8 applications give developers a supported path to bring DLSS-enabled projects to RTX Spark systems, according to NVIDIA.

Nemotron 3.5 Bolt for AI characters

NVIDIA ACE is the company's suite for building in-game AI agents. NVIDIA points to PUBG, which rolled out AI teammates in a public beta powered by its first-generation Nemotron small language model, and to Aniimo, which uses a second-generation model for voice-responsive AI pets and is due in early 2027.

Nemotron 3.5 Bolt E4B 4GB is now available. NVIDIA claims it offers about 40% higher average quality than its prior model on relevant benchmarks and outputs up to about 80% fewer tokens than peers. It runs locally on GeForce RTX and RTX Spark and is designed to leave headroom for graphics rendering and local speech models.

• Multi-turn tool use: calls application APIs across a conversation.
• Conversational mode: encourages concise, speech-ready responses.
• Configurable reasoning: can be enabled for complex requests or disabled for speed.
• 64k context: holds history, instructions, tool definitions, retrieved information and game state.
• Multilingual support: 35 languages.

Developers can get the model through the In-Game Inferencing (IGI) SDK. To try it in a game, NVIDIA suggests SkyrimNET, a modding platform by MinLL that turns NPCs in The Elder Scrolls V: Skyrim into interactive characters with memory, voice and autonomy.

RTX Kit, Nsight and what it means for developers

Eight RTX Kit technologies are now published with RTX Spark support. They include RTX Neural Texture Compression, which uses AI to shrink texture storage while keeping visual fidelity comparable to traditional block compression, as well as RTX Mega Geometry and RTX Character Rendering for strand-based hair and skin with path tracing.

On tooling, NVIDIA says its Nsight tools support native Arm64, emulated x64 and Arm64EC applications on RTX Spark from day one. Nsight Graphics 2026.3 adds a Windows on Arm installer, and Nsight Systems 2026.5 offers a single installer for all Windows versions. The NVIDIA RTX Branch of Unreal Engine 5.8 provides an updated base for ray tracing and neural graphics, and MotionBricks adds new SOMA human-motion models, downloadable checkpoints and ONNX exports.

Taken together, the announcements give studios a route to bring DLSS 5, local AI characters and neural rendering to Arm-based RTX Spark systems, with profiling and engine support available from launch.