开源模型性能逼近闭源，成本却低一个数量级——做智能体或编程的开发者可以直接在 Hugging Face 上试用，省下不少推理预算。
NVIDIA 发布了开源模型 Nemotron 3 Ultra，专为智能体任务和编程设计。该模型在性能上接近闭源模型 GPT 5.5，但推理成本仅为后者的十分之一（$0.051 vs $0.57）。Nemotron 3 Ultra 速度比同类模型快5倍，成本低30%，已在 Hugging Face 上开源。这标志着开源模型与闭源模型之间的质量差距正在迅速缩小，尤其适合处理大型代码库和智能体应用。
This NVIDIA model is ridiculously good Same prompt, 10x cheaper. Nemotron 3 Ultra: $0.051 GPT 5.5...
This NVIDIA model is ridiculously good Same prompt, 10x cheaper. Nemotron 3 Ultra: $0.051 GPT 5.5: $0.57 The quality difference between open source and closed models is way smaller than you think. Your browser does not support the video tag. 🔗 View on Twitter Paul Couvert @itsPaulAi Nvidia has just released a REALLY powerful open AI model 🔥 Nemotron 3 Ultra is tailor-made for agentic tasks and coding while being: - 5x faster than other models - Up to 30% cheaper - Available on Hugging Face Perfect for large codebases, Hermes Agent and more! Your browser does not support the video tag. 🔗 View on Twitter 🔗 View Quoted Tweet 💬 3 🔄 2 ❤️ 7 👀 707 📊 3 ⚡