📄 中文摘要
AI 交互的核心是令牌，涵盖医疗诊断、互动游戏角色对话和客户服务代理的自主解决方案。为了扩大这些 AI 交互的规模，企业需要考虑是否能够承担更多的令牌费用。通过优化令牌经济学，企业可以显著降低成本，从而实现更高效的 AI 应用。这种方法不仅提升了 AI 模型的可负担性，还促进了开源模型的广泛应用，推动了行业的创新与发展。
领先的推理服务提供商通过 NVIDIA Blackwell 上的开源模型将 AI 成本降低至 10 倍
📄 English Summary
Leading Inference Providers Cut AI Costs by up to 10x With Open Source Models on NVIDIA Blackwell
The core of AI interactions lies in tokens, which are essential for applications ranging from diagnostic insights in healthcare to character dialogues in interactive games and autonomous resolutions from customer service agents. To scale these AI interactions, businesses must evaluate their ability to afford more tokens. By improving tokenomics, companies can significantly reduce costs, enabling more efficient AI applications. This approach not only enhances the affordability of AI models but also encourages the widespread adoption of open-source models, driving innovation and growth within the industry.