Tag: muse-glimmer
-
NVIDIA and open source community advance local AI models and agents
605BNVIDIA is celebrating open source communities and partners building local AI throughout August, highlighting new models and tools for developers. The company released Nemotron 3.5 Lightning, a 30B mixture-of-experts model delivering up to 4x faster token generation than comparable open models, which runs on RTX PCs, DGX Spark, and Jetson systems. NVIDIA also announced support for Qwen3.8-27B reaching 131 tokens per second on a GeForce RTX 5090, and highlighted new models from partners including Meta's Muse Glimmer, LTX-2.5 video generation model, and updates to DGX Spark clustering via NVIDIA Sync.