Which AI models does NEXSTA use?
The roles of OpenAI and Z.ai in NEXSTA, why model updates are tested and what should remain reliable for website owners.
The model inside an AI employee
NEXSTA’s global configuration uses OpenAI’s GPT-5.6 Sol; its Russian configuration uses Z.ai’s GLM family. Choosing a supplier by region does not change the product’s purpose: the AI employee uses company information to answer website visitors and help pass inquiries to the owner.
Product settings should not be confused with official model names. This article does not claim unverified GLM versions, API prices or benchmark results. Supplier documentation describes the model families, not the configuration of an individual NEXSTA account.
Why the model stack changes
Suppliers release new versions and change availability or terms. An update is useful when it improves relevant answers while keeping response times acceptable. A newer version number alone is not a reason to switch.
Before deployment, compare answers to real website questions, responses to unknown facts and adherence to the assigned role. An unjustifiably confident answer is worse than an honest offer to check with a specialist.
What matters to a business owner
NEXSTA evaluates stable new model versions before using them. Company materials remain the source of facts, and the owner continues to handle inquiries in the dashboard.
A model announcement in the blog does not mean that it has been enabled in the product. Actual NEXSTA changes appear separately in updates. To assess value, test answers to your own questions instead of choosing an AI employee only by the model name.
Continue reading: GPT-5.6 and business. Capabilities of AI employees - On the NEXSTA website.
Official source, checked on September 16, 2026: OpenAI: GPT-5.6