Connect with us

Hi, what are you looking for?

AI Technology

NVIDIA Reveals ComfyUI Update, Enhancing Local AI Video Generation with 2.5x Performance Boost

NVIDIA unveils ComfyUI update with 2.5x performance boost for local AI video generation on RTX GPUs, streamlining workflows for artists and developers.

Game developers and artists are raising the bar for immersive experiences on NVIDIA RTX AI PCs, building cinematic worlds and iconic characters. At the Game Developers Conference (GDC) held in San Francisco this week, NVIDIA unveiled a suite of updates aimed at enhancing the capabilities of AI video generation for concept development and storyboarding, particularly on its RTX GPUs and the NVIDIA DGX Spark desktop supercomputer.

Among the notable announcements were ComfyUI’s new App View, which features a simplified interface designed to lower the barrier for entry for the popular generative AI tool. The updates also include the introduction of RTX Video Super Resolution, a real-time 4K upscaler that is ideal for video generation and is now available as a Python Wheel for developers. Furthermore, NVIDIA introduced NVFP4 and FP8 model variants for FLUX.2 Klein, delivering performance gains of up to 2.5x and reducing memory usage by 60% for both models.

ComfyUI’s new App View mode streamlines the user experience, making it easier for beginners to experiment with state-of-the-art AI models directly on their laptops or desktops. Users can now enter a prompt, adjust basic parameters, and generate outputs without needing to navigate complex node graphs. For those familiar with node-based workflows, the traditional Node View remains available, allowing seamless transitions between the two modes. Performance enhancements for RTX GPUs have been significant, showing an increase of 40% since September, while ComfyUI now supports NVFP4 and FP8 data formats natively. Collectively, these improvements enable performance that is 2.5x faster with NVIDIA GeForce RTX 50 Series GPUs’ NVFP4 format and 1.7x faster with FP8, with VRAM usage reduced by 60% and 40% respectively.

NVIDIA had previously announced several models with NVFP4 and FP8 support earlier this year at CES, and additional models such as LTX-2.3 are being rolled out with NVFP4 support arriving soon. Users can easily get started by downloading the NVFP4 and FP8 checkpoints from Hugging Face, loading the default workflows in ComfyUI via the Template Browser, and replacing the standard model checkpoint with the newly downloaded one. App View mode is available now, providing an accessible entry point for users.

In another significant development, NVIDIA introduced RTX Video Super Resolution, which allows users to upscale generated video to 4K quality with remarkable speed. Traditionally, artists would need to generate smaller preview clips and upscale them through lengthy processes. The new RTX Video functionality, available as a node for ComfyUI, can be utilized to build customized video workflows. For AI developers, NVIDIA has released a free Python package via the PyPI repository, complete with sample code on GitHub. This package allows programmatic access to the same AI upscaling technology that powers RTX Video, operating directly on RTX GPU Tensor Cores to provide 4K upscaling that is 30x faster than widely used local upscalers, while consuming significantly less VRAM.

As the AI landscape continues to evolve, NVIDIA’s innovations are aimed at simplifying the creative process for artists and developers alike. The introduction of tools that enhance performance and streamline workflows underlines the company’s commitment to improving the user experience. Attendees at GDC can explore these advancements further, with NVIDIA hosting training sessions aimed at helping users build RTX-accelerated generative workflows for various applications, including images and video.

In addition to these updates, NVIDIA is set to participate in GTC from March 16-19 in San Jose, showcasing further innovations and providing training on new tools. Other updates of interest include the LTX Desktop, a fully local, open-source video editor optimized for NVIDIA GPUs; LM Link, which allows devices running LM Studio to connect remotely; and a new NVIDIA RTX Remix update that will enhance modding capabilities with advanced particle effects.

As the industry moves toward increasingly sophisticated AI applications, NVIDIA remains at the forefront, continually enhancing the hardware and software that power modern creative workflows. With these latest announcements, the company reinforces its role as a critical player in the intersection of gaming, video production, and artificial intelligence, setting the stage for even more groundbreaking developments in the future.

Staff
Written By

The AiPressa Staff team brings you comprehensive coverage of the artificial intelligence industry, including breaking news, research developments, business trends, and policy updates. Our mission is to keep you informed about the rapidly evolving world of AI technology.

You May Also Like

AI Finance

C3.ai's stock plunges as billings decline for four consecutive quarters, raising concerns about cash burn and investor confidence in 2026.

AI Tools

Nvidia unveils open-source AI platform 'NemoClaw' ahead of GTC 2026, amplifying its revenue potential and securing a $2B partnership with Thinking Machines Lab.

AI Regulation

Around 10,000 writers, led by Ed Newton-Rex, protest AI copyright threats at the London Book Fair as publishers propose a collective licensing scheme.

AI Research

Microsoft appoints Peter Lee as President of Microsoft Science to drive AI integration in research, aiming to transform biomedical sciences and enhance discovery.

AI Marketing

Buildout Inc. launches Buildout CRM, unifying commercial real estate workflows into an AI-powered platform, boosting operational efficiency and reducing deal costs.

AI Generative

Meta's Oversight Board demands a major overhaul of its inadequate AI deepfake detection systems to combat rising misinformation during critical global events.

Top Stories

Google.org partners with CDP to develop an AI tool that optimizes environmental data usage, empowering cities to combat climate risks effectively.

AI Cybersecurity

University of Surrey's TwinGuard AI system neutralises 5G cyber-attacks in under 100 milliseconds, enhancing future network security against evolving threats.

© 2025 AIPressa · Part of Buzzora Media · All rights reserved. This website provides general news and educational content for informational purposes only. While we strive for accuracy, we do not guarantee the completeness or reliability of the information presented. The content should not be considered professional advice of any kind. Readers are encouraged to verify facts and consult appropriate experts when needed. We are not responsible for any loss or inconvenience resulting from the use of information on this site. Some images used on this website are generated with artificial intelligence and are illustrative in nature. They may not accurately represent the products, people, or events described in the articles.