Contextual AI’s new AI model crushes GPT-4o in accuracy — here’s why it matters
1 min read
Summary
Contextual AI, a startup founded by the pioneers of retrieval-augmented generation (RAG) technology, has introduced its grounded language model (GLM), which it claims is the most factually accurate in the industry, outperforming the likes of Google, Anthropic and OpenAI on the FACTS benchmark for truthfulness.
RAG is a technique that optimises all system components to deliver the most relevant and accurate response in an enterprise RAG application, where precision is a priority.
While general-purpose models are designed for flexibility, high-stakes enterprise environments require AI tools that are focused on factual precision, able to provide accurate information or state when it doesn’t know something.
Contextual AI’s platform supports structured and unstructured data and plans to release a specialised re-ranker and expanded document-understanding capabilities in the near future, along with more agentic capabilities.