Contextual AI’s new AI model crushes GPT-4o in accuracy — here’s why it matters

Contextual AI unveiled its grounded language model (GLM) at the moment, claiming it delivers the best factual accuracy within the {industry} by outperforming main AI methods from Google, Anthropic and OpenAI on a key benchmark for truthfulness.

The startup, based by the pioneers of retrieval-augmented generation (RAG) expertise, reported that its GLM achieved an 88% factuality rating on the FACTS benchmark, in comparison with 84.6% for Google’s Gemini 2.0 Flash, 79.4% for Anthropic’s Claude 3.5 Sonnet and 78.8% for OpenAI’s GPT-4o.

Whereas giant language fashions have remodeled enterprise software program, factual inaccuracies — typically referred to as hallucinations — stay a vital problem for enterprise adoption. Contextual AI goals to unravel this by making a mannequin particularly optimized for enterprise RAG purposes the place accuracy is paramount.

“We knew that a part of the answer could be a way referred to as RAG — retrieval-augmented technology,” mentioned Douwe Kiela, CEO and cofounder of Contextual AI, in an unique interview with VentureBeat. “And we knew that as a result of RAG is initially my concept. What this firm is about is de facto about doing RAG the fitting manner, to sort of the subsequent degree of doing RAG.”

The corporate’s focus differs considerably from general-purpose fashions like ChatGPT or Claude, that are designed to deal with every little thing from inventive writing to technical documentation. Contextual AI as an alternative targets high-stakes enterprise environments the place factual precision outweighs inventive flexibility.

“In case you have a RAG drawback and also you’re in an enterprise setting in a extremely regulated {industry}, you don’t have any tolerance in any way for hallucination,” defined Kiela. “The identical general-purpose language mannequin that’s helpful for the advertising division shouldn’t be what you need in an enterprise setting the place you might be way more delicate to errors.”

How Contextual AI makes ‘groundedness’ the brand new gold customary for enterprise language fashions

The idea of “groundedness” — making certain AI responses stick strictly to info explicitly supplied within the context — has emerged as a vital requirement for enterprise AI methods. In regulated industries like finance, healthcare and telecommunications, corporations want AI that both delivers correct info or explicitly acknowledges when it doesn’t know one thing.

Kiela supplied an instance of how this strict groundedness works: “If you happen to give a recipe or a components to a regular language mannequin, and someplace in it, you say, ‘however that is solely true for many instances,’ most language fashions are nonetheless simply going to provide the recipe assuming it’s true. However our language mannequin says, ‘Really, it solely says that that is true for many instances.’ It’s capturing this extra little bit of nuance.”

The power to say “I don’t know” is an important one for enterprise settings. “Which is mostly a very highly effective function, if you consider it in an enterprise setting,” Kiela added.

Contextual AI’s RAG 2.0: A extra built-in method to course of firm info

Contextual AI’s platform is constructed on what it calls “RAG 2.0,” an strategy that strikes past merely connecting off-the-shelf parts.

“A typical RAG system makes use of a frozen off-the-shelf mannequin for embeddings, a vector database for retrieval, and a black-box language mannequin for technology, stitched collectively via prompting or an orchestration framework,” in line with an organization assertion. “This results in a ‘Frankenstein’s monster’ of generative AI: the person parts technically work, however the entire is way from optimum.”

As a substitute, Contextual AI collectively optimizes all parts of the system. “We’ve got this mixture-of-retrievers element, which is mostly a method to do clever retrieval,” Kiela defined. “It seems on the query, after which it thinks, basically, like many of the newest technology of fashions, it thinks, [and] first it plans a technique for doing a retrieval.”

This whole system works in coordination with what Kiela calls “one of the best re-ranker on the earth,” which helps prioritize probably the most related info earlier than sending it to the grounded language mannequin.

Past plain textual content: Contextual AI now reads charts and connects to databases

Whereas the newly introduced GLM focuses on textual content technology, Contextual AI’s platform has not too long ago added assist for multimodal content material together with charts, diagrams and structured information from fashionable platforms like BigQuery, Snowflake, Redshift and Postgres.

“Essentially the most difficult issues in enterprises are on the intersection of unstructured and structured information,” Kiela famous. “What I’m principally enthusiastic about is de facto this intersection of structured and unstructured information. A lot of the actually thrilling issues in giant enterprises are smack bang on the intersection of structured and unstructured, the place you might have some database information, some transactions, perhaps some coverage paperwork, perhaps a bunch of different issues.”

The platform already helps a wide range of complicated visualizations, together with circuit diagrams within the semiconductor {industry}, in line with Kiela.

Contextual AI’s future plans: Creating extra dependable instruments for on a regular basis enterprise

Contextual AI plans to launch its specialised re-ranker element shortly after the GLM launch, adopted by expanded document-understanding capabilities. The corporate additionally has experimental options for extra agentic capabilities in growth.

Based in 2023 by Kiela and Amanpreet Singh, who beforehand labored at Meta’s Elementary AI Analysis (FAIR) staff and Hugging Face, Contextual AI has secured prospects together with HSBC, Qualcomm and the Economist. The corporate positions itself as serving to enterprises lastly notice concrete returns on their AI investments.

“That is actually a possibility for corporations who’re perhaps below strain to start out delivering ROI from AI to start out taking a look at extra specialised options that really remedy their issues,” Kiela mentioned. “And a part of that basically is having a grounded language mannequin that’s perhaps a bit extra boring than a regular language mannequin, but it surely’s actually good at ensuring that it’s grounded within the context and you can actually belief it to do its job.”