NVIDIA Releases Free Open AI Models and Enterprise Tools
TechNews reported that NVIDIA announced a set of free open AI models and tools on November 5 for scientific research, engineering analysis and enterprise applications. Nemotron 3.5 Super VL is designed to analyze screenshots, PDFs and charts, and the report says it offers faster performance than its predecessor. It is available for free through OpenCode, OpenRouter, Hugging Face, Ollama and LM Studio. NVIDIA also expanded its open-source Agent Toolkit with Knowledge Foundation, which allows enterprise AI agents to use internal data for reasoning without moving company data and while retaining control over access and actions. The Kumo model family covers tabular, relational, time-series and predictive workloads.
The release broadens NVIDIA’s free open-model offering across visual analysis, enterprise knowledge-grounded agents and structured-data workloads. The report also says free access through the listed platforms lowers the barriers to testing and deployment.
- )
輝達（NVIDIA）11 月 5 日宣布推出一系列免費的開放 AI 模型與工具，目標是協助開發者更快將人工智慧導入科學研究、工程分析與企業應用。這批模型與工具涵蓋不同工作負載，從視覺理解到結構化資料推理，顯示 NVIDIA 持續擴大其開放模型布局。
在視覺模型方面，NVIDIA 釋出的 Nemotron 3.5 Super VL 可用於分析截圖、PDF 與圖表，主打更快的運算速度與優於前代的表現。根據相關資訊，這款模型可透過 OpenCode、OpenRouter、Hugging Face、Ollama 與 LM Studio 免費使用，降低開發者測試與部署的門檻。
除了視覺能力，NVIDIA 也同步擴充其開源 Agent Toolkit，新增 Knowledge Foundation，讓企業 AI 代理在不搬移公司資料的情況下，直接以內部資料做為推理基礎，同時維持對存取與行動的控制。這項更新被視為企業導入代理式 AI 的重要基礎設施。
#NVIDIA#Nemotron 3.5 Super VL#Agent Toolkit#Knowledge Foundation#Kumo#OpenCode#OpenRouter#Hugging Face#Ollama#LM Studio