Applied research for production AI
Our research starts from a simple belief: the most capable AI systems are those that deeply understand context. Our research team pushes the frontier of AI, with state-of-the-art models that turn raw enterprise data into real intelligence you can trust—in production, at scale.
Featured Research
All Research
Introducing LMUnit: Natural language unit testing for LLM evaluation
Contextual AI Team
December 18, 2024
Introducing OLMoE: A small, fully open, and state-of-the-art mixture-of-experts LLM
Niklas Muennighoff
September 4, 2024
Addressing Underspecification in Language Model Alignment
Karel D’Oosterlinck
August 13, 2024
Introducing RAG 2.0
Contextual AI Team
March 19, 2024
Training with GRIT for better LLMs and faster RAG
Niklas Muennighoff
February 15, 2024
Better, Cheaper, Faster LLM Alignment with KTO
Kawin Ethayarajh
December 7, 2023
Vision Augmented Language Models: Computer vision through the LENS of natural language
William Berrios
June 27, 2023