Apollo-1

Apollo-1 is a neuro-symbolic foundation model that combines neural learning with symbolic reasoning to create reliable AI agents for enterprises. Unlike traditional LLMs, Apollo-1 offers a white-box view of its reasoning, ensuring traceable decision-making. It is tool-native, capable of activating any number of tools (APIs, CRM, inventory systems) with high accuracy. The model supports rule-based controllability, allowing companies to enforce business rules, preferences, and policies. Apollo-1 provides grounded responses by incorporating context and sources, and it continuously improves through automatic fine-tuning from human feedback. Deployed across regulated and unregulated industries, Apollo-1 is designed for enterprise use cases requiring predictability and control. General availability is scheduled for Q2 2026.