Intel has released a public beta of its enterprise hybrid AI platform SuperClo based on the open-source OpenClo framework, HotHardware reported. The platform dynamically distributes work between local systems and the cloud. High-frequency tasks such as file parsing, memory search and masking sensitive data run on-premises, while complex multi-step reasoning and deep web searches are selectively sent to cloud large language models. Intel said this can cut cloud token consumption by over 70 percent.

Companies are moving away from simply increasing artificial intelligence use and are adopting cost-control strategies that assign different AI models by task. The approach, known as “model mixing,” is gaining attention as AI spending rises. Executives and users describe choosing cheaper models for routine work and reserving top-tier systems for complex tasks. Model-routing software that automatically assigns requests to suitable models is also drawing interest, as firms seek to balance performance and costs.