Nvidia releases Nemotron 3.5 Super VL, a 120B open-weight multimodal model
Nemotron 3.5 Super VL has 120B parameters with only 12B active per token, targets reasoning, tool use, images and PDFs, and Nvidia claims up to 5x faster throughput than comparable open models.
🚨 GOOD NEWS : $NVDA RELEASES A 120B OPEN-WEIGHT AI MODEL. AND $PLTR IS ONE OF THE BIG BENEFICIARIES. Today, NVIDIA launched Nemotron 3.5 Super VL, a 120B-parameter multimodal model with only 12B parameters active per token. NVIDIA claims up to 5x faster throughput than comparable open models, with support for reasoning, tool use, images, PDFs and complex agent workflows. Now remember, $PLTR and NVIDIA already have a partnership bringing Nemotron models into Foundry and AIP, grounded in Palantir's Ontology. Imagine a manufacturer using Nemotron to analyze inspection images, maintenance reports and equipment documentation. Palantir's Ontology connects those findings to actual equipment, inventory, production schedules and business rules. AIP helps turn the analysis into operational workflows with the right permissions and approvals. And because Nemotron is open-weight, customers can customize and deploy it inside their own infrastructure. For $PLTR, more capable and cheaper open models could make AIP deployments more attractive without Palantir having to develop the foundation models itself. The existing NVIDIA partnership is confirmed, although this specific Super VL model hasn't been announced as an AIP integration. NVIDIA will keep improving the models, while Palantir will help enterprises put that intelligence to work inside their operations.