Questions
Intelligence Stack questions
The layers, ZB Model, the network and ZB Compute.
What are the seven layers?
L7 Maestro, L6 the Harness, L5 ZB Model, L4 the Network of Intelligence, L3 the Middle Layer, L2 Security and Confidential Computing, and L1 ZB Compute. See the stack in one view.
What is the difference between the Harness and ZB Model?
The Harness runs the work: it keeps many models and agents working from one context until the plan is complete. ZB Model decides the work: it plans, routes each task and checks each result. See How the layers work together.
How is ZB Model different from a model router?
Model routing chooses where a single request goes. ZB Model sees the whole job, from the plan through the checked result, and routing is one of the decisions it makes along the way.
Which models are in the network?
Frontier models for hard reasoning, fast models for simple tasks, open models that run on ZB Compute, and specialist agents for coding, design, testing and research, from every major AI provider. See the Network of Intelligence.
What happens when a better model is released?
It joins the network on the day it is released. ZB Model learns what it is best at and starts routing those tasks to it.
Can one project use both open and frontier models?
Yes. ZB Model can coordinate open-model inference on ZB Compute with calls to frontier models through the providers that host them.
What does the Middle Layer do?
It gives every model and agent the same way to reach memory, tools, data and compute, connects your project to the systems your business already uses, and records which model did each task. See the Middle Layer.
What is ZB Compute?
The compute infrastructure that runs the models, the agents and your finished solution. We are starting with 64 GPUs, in a cloud or inside a country's border. See ZB Compute.
What does serverless mean on ZB Compute?
ZB Compute manages the compute for inference and agent workloads. You bring the job, and ZB Compute provides the environment that runs it.