NVIDIA announced on October 13, 2026, that its ecosystem partners have prepared software updates and new features for RTX Spark laptops, which are scheduled to launch on October 16, 2026, according to a joint announcement by NVIDIA and Microsoft published October 7, 2026. The updates span AI agent applications, creative tools, developer frameworks, and gaming software.
AI Agent Software and Models
NVIDIA stated that agent applications including GitHub Copilot, Hermes, OpenClaw, Perplexity, Pi, and WorkBuddy will streamline setup on RTX Spark and add security features through OpenShell and Microsoft Execution Containers (MXC). The company said these apps will recommend optimized local models and configurations.
NVIDIA Nemotron 3.5 Bolt 74B-A9B 24GB, described by the company as purpose-built for agentic AI on RTX Spark, is scheduled for release on October 15, 2026. NVIDIA characterized the model as a 74-billion-parameter, 9-billion-active-parameter mixture-of-experts model compressed to 24 GB using 2-bit quantization. The company stated the model will be available through llama.cpp, vLLM, Ollama, Unsloth, and LM Studio at launch, and accessible via agents such as Hermes and OpenClaw. NVIDIA said Nemotron 3.5 Bolt 74B will be fully open source, including training data, weights, and recipes.
NVIDIA PAIR (Personal AI Router) is moving out of beta as of October 13, 2026. The free, open-source tool is designed to distribute AI workloads across compatible PCs on a local network. The official release adds llama.cpp support, direct model downloads from Hugging Face, and RTX Spark compatibility, according to NVIDIA.
Creative Application Optimizations
Adobe is optimizing Photoshop, Premiere, and Substance 3D for RTX Spark, with NVIDIA stating the updates will deliver up to 2x faster performance across AI, editing, color, and effects workflows. NVIDIA said GPU-accelerated compositing in Photoshop will make complex image editing faster, while Premiere will accelerate playback, rendering, color grading, and AI features such as Enhance Speech and Object Mask. AI feature optimizations in Photoshop and Premiere are available in beta as of October 13, 2026, with Substance 3D updates coming later, according to the announcement.
NVIDIA Broadcast 2.3, available for download on October 13, 2026, adds RTX Video Super Resolution and Frame Generation AI effects, Studio Voice profiles, and an Offline File Processor for applying effects to prerecorded content. The update includes a Model Context Protocol (MCP) server that allows local AI agents to adjust effects and process files through the Broadcast app, NVIDIA said.
ComfyUI is adding RTX Video Frame Generation and RTX Video HDR nodes, plus EXR file export capability. NVIDIA stated that RTX Video Frame Generation increases video frame rates for GeForce RTX 40 Series GPUs and newer, while RTX Video HDR converts standard dynamic range footage to high dynamic range.
Blender version 5.3, scheduled for November 2026, will add support for DLSS Ray Reconstruction to improve viewport interactivity, according to NVIDIA. A beta version with RTX Spark compatibility is available as of October 13, 2026.
Developer Tools and Frameworks
NVIDIA said it has partnered with Microsoft to prepare developer tools for RTX Spark. The announcement listed the first PyTorch-CUDA build for Windows on Arm, the first Python build native by default on Windows on Arm with Python 3.15, and the first native Conda installer for Windows on Arm with over 2,000 Conda packages. NVIDIA stated that 97% of the top 15,000 Python wheels are ready for the platform.
GitHub Copilot now supports MXC and is updating its HydraFusion model router to support local models, enabling developers to reduce cloud token consumption while accessing frontier intelligence, according to the announcement. Microsoft is making coding models such as MAI 1.1 Coding Flash available locally and adding llama.cpp as a backend to Windows ML, NVIDIA said.
Llama.cpp is releasing Llama Windows, a Windows-native companion app for llama.cpp, on October 13, 2026. The app provides a user interface to download and run models locally without command-line interaction. Llama.cpp now supports decision models on RTX Spark and GeForce RTX GPUs, which NVIDIA described as enabling AI agents to make decisions up to 200x faster than traditional agents.
Gaming Features and Anti-Cheat Support
NVIDIA stated that hundreds of games will be available when RTX Spark laptops launch on October 16, 2026, including CONTROL Resonant, Cyberpunk 2077, Fortnite, Gears of War: E-Day, Genshin Impact, Hogwarts Legacy, Honkai: Star Rail, Marvel Rivals, NARAKA: BLADEPOINT, NBA 2K27, and World of Warcraft. The company said it has worked with game developers to enable anti-cheat support for popular multiplayer games.
A new, 2x faster DLSS 5 AI model debuts in NBA 2K27 on October 16, 2026, with NVIDIA claiming frame rate boosts of up to 1.4x. The faster model will also power DLSS 5 in Hogwarts Legacy starting November 5, 2026, and Where Winds Meet in late November 2026, according to the announcement. DLSS 5 is coming to NVIDIA RTX Remix on October 23, 2026.
Nemotron 3.5 Bolt E4B 4GB, described by NVIDIA as the third-generation small language model for NVIDIA ACE AI game characters, is scheduled for release on October 23, 2026. The model is designed for low memory footprint and fast time to first token, enabling responsive local experiences for interactive characters and embedded interfaces. Starting October 23, 2026, the model will be available in SkyrimNET, an AI NPC modding platform by MinLL for The Elder Scrolls V: Skyrim.
RTX Remix 1.6, launching October 23, 2026, includes a major overhaul to AI-powered texture generation workflows with ComfyUI Version 2, which replaces REST API node connections with a dedicated AI tools layout in the sidebar, according to NVIDIA.
Platform Specifications
According to NVIDIA documentation verified as of October 13, 2026, RTX Spark N1X is an Arm-based computing platform combining a Grace CPU and Blackwell RTX GPU for Windows 11 laptops and compact desktops. NVIDIA lists two laptop configurations: 6,144 GPU cores with 20 CPU cores and up to 128 GB LPDDR5X unified memory, or 5,120 GPU cores with 18 CPU cores and up to 64 GB LPDDR5X, both with a 45–80 W TDP. The desktop configuration lists 6,144 GPU cores, 20 CPU cores, up to 128 GB LPDDR5X, and a 140 W TDP. NVIDIA advertises up to 1 petaflop of FP4 AI performance and local execution of models up to 120 billion parameters with a 1-million-token context; these are vendor-stated capabilities.
Source: blogs.nvidia.com.
Be the first to comment