NVIDIA has partnered with Black Forest Labs to optimize the 'FLUX.1 Kontext' model for RTX GPUs to maximize the efficiency and quality of AI image generation.

This collaboration is gaining attention as a key technology that enables high-quality visuals and a real-time editing environment while simplifying the image creation workflow.

Flux.1 Context is a context-based open-weight model that supports both text and image prompts, enabling image creation and editing using only intuitive natural language input.

It features flexible style changes and precise detail adjustments based on reference images without the need for complex masks, depth maps, or edge maps.

NVIDIA has significantly reduced VRAM requirements by optimizing models using TensorRT and quantization techniques.

This enables reliable implementation of fast inference and iterative editing on RTX GPU-based devices, and is available out of the box in ComfyUI and Playground.

Additionally, a version of NVIDIA NIM microservices is scheduled to be released in August.

The model's flexibility and controllability are highlighted by its features:

Character consistency is It maintains its unique characteristics even in different scenes and angles, and custom editing can selectively modify only some elements of the image. Style switching can apply the mood of a reference image to a new scene, and is performed in real time, enabling fast feedback and iterative creation with short delays.

Model weights and TensorRT variants are available for free download via Hugging Face, and the Ada-based FP8 is down to 12 GB, and the Blackwell-based FP4 is down to 7 GB, ensuring accessibility for running in local environments.

SVDQuant technology, specifically optimized for the NVIDIA GeForce RTX 40/50 series, provides efficient processing speed while maintaining high image quality.

In addition to AI image generation, Google's recently released small language model 'Gemma 3n' also achieves optimal performance on NVIDIA RTX and Jetson platforms.

Gemma 3n works with Ollama and Llama.cpp-based applications to provide RTX-accelerated natural language processing capabilities, opening up a wide range of opportunities for AI enthusiasts and developers.

In addition, NVIDIA is hosting the 'Project G-Assist Plugin Hackathon' online until July 17th to provide developers with the opportunity to create custom AI plugins.

A G-Assist webinar is also scheduled for July 10th, where you can exchange various information through real-time Q&A.

NVIDIA is committed to community-driven technology sharing and content creation through its Discord server and RTX AI Garage blog series, as well as providing ongoing information on the next generation of creative innovations with AI PCs and workstations.

Flux.1 Context and NVIDIA's latest AI technologies can be found on NVIDIA's official blog and social media channels, as well as Hugging Face. You can also get the latest news through our related newsletters.