NVIDIA ACE has added the open-source Qwen3 SLM. The company has also announced the new Bonsai Dioramas and IGI SDK updates.
NVIDIA ACE now supports the open-source Qwen3-8B small language model (SLM), allowing developers to deploy dynamic, real-time NPC characters directly in on-device PC games. The open-source Qwen3-8B AI model is now available through NVIDIA ACE as an In-Game Inferencing (IGI) SDK plugin, simplifying integration within the pipeline and optimizing simultaneous AI inference and graphics processing for accelerated game performance.
Qwen3 enables developers to create more advanced in-game characters capable of real-time reasoning, dynamic responses to player input, and contextually accurate actions — all without relying on predefined scripts.
A series of new updates have been released for Unreal Engine’s NVIDIA RTX branch (NvRTX). Among these are Unreal Engine 5.6.1, RTX Mega Geometry, and several ReSTIR PT performance and image quality improvements.
The Bonsai Diorama demo accompanies the NvRTX 5.6.1 release, showcasing RTX Mega Geometry while also leveraging ReSTIR PT and the DLSS 4 technology suite.
Several enhancements have been introduced to the core IGI SDK.
Magpie Flow has been added to enable real-time, on-device text-to-speech with multilingual support for Spanish and German. MultiLoRA adapters have been introduced to allow fast fine-tuning of model weights with minimal computational load.
Additionally, CUDA support for Vulkan backends has been implemented to optimize graphics and AI workloads, delivering accelerated performance.
