[NVIDIA Launches Nemotron 3.5 Lightning Open Model, Supporting Local Deployment on RTX PCs and Other Devices] According to Jin10, NVIDIA has launched the Nemotron 3.5 Lightning open model, which can run locally on NVIDIA RTX personal computers, DGX Spark, OEM systems equipped with GB10, and the Jetson platform. NVIDIA has partnered with vLLM, Ollama, llama.cpp, LM Studio, and Unsloth to provide support for this model. NVIDIA also stated that Nemo Switchyard can reduce the cost of completing benchmark tests to one-third of Opus 4.8. Additionally, according to The Information, citing NVIDIA employees involved in the project’s development, NVIDIA is working on the Nemotron 4 open-source AI model, with its largest version expected to feature at least 1 trillion parameters.