Skip to main content

​
Overview

By accessing Contextual AI’s Generate API component, you can generate responses using Contextual AI’s Grounded Language Model (GLM), an LLM engineered specifically to prioritize faithfulness to in-context retrievals over parametric knowledge to reduce hallucinations in retrieval-augmented generation (RAG) and agentic use cases.

​
Key Features

  • State-of-the-art groundedness with top performance on FACTS and real enterprise benchmarks.
  • Hallucination-resistant generation that prioritizes retrieved knowledge over training data.
  • Inline source attributions for transparent, traceable responses.
  • Reliable multi-turn reasoning that maintains grounding in long or complex workflows.
  • RAG-optimized design built specifically for retrieval and agentic applications.
  • Configurable groundedness via the avoid_commentary flag.
  • Easy integration through /generate, Python SDK, and LangChain examples.
  • Enhanced performance when used with the Contextual AI Platform’s unified RAG stack.

​
Getting Started

See the Generate How-to guide for a detailed walkthrough on how to use the Generate API.

​
Additional Resources