You are currently viewing OpenPipe Secures $6.7M to Support Businesses in Curbing LLM Model costs
Citation: Image used for information purpose only. Picture Credit:

OpenPipe Secures $6.7M to Support Businesses in Curbing LLM Model costs

GeekWire’s in-depth startup coverage highlights the Pacific Northwest entrepreneurial ecosystem.

OpenPipe, a Seattle startup that promises to make it faster and cheaper for businesses to train and deploy massive language models, has announced a $6.7 million seed round.
The firm, which participated in Y Combinator’s summer batch last year, allows developers to create LLM models that are personalized to their specific use case. The notion is that smaller models can provide greater performance at a lesser cost than bigger general-purpose models like OpenAI’s ChatGPT.

OpenPipe believes that firms seeking to use generative AI may not want a chatbot that can answer general questions, but rather one that is well-versed in their product line and company policy to offer customer service or other services.
For example, OpenPipe mentioned a financial services organization that was utilizing OpenAi to process call transcripts and extract information like as credit card balances. It was able to considerably cut costs and errors after transitioning to OpenPipe.

“Customers end up paying much less than their previous OpenAI bill while benefiting from the higher-quality responses a model tailored to their specific use case produces,” said OpenPipe CEO Kyle Corbitt.

OpenPipe debuted last year and reported “significant growth,” but declined to provide revenue figures. It generates revenue by charging customers for both fine-tuning their models and using them in production.
Companies across industries are increasing their usage of generative AI, but the expensive cost of training, deploying, and updating models remains a major problem.

“We’ve found that in practice most businesses are using generative AI for specific narrow use cases within their product or service, which is a great fit for these smaller specialized models,” he said.
OpenPipe’s software does not require advanced machine learning or data science expertise, which is one of its selling points.
“Our users are fullstack app engineers, and they’re able to successfully train really strong models for their specific use cases without involving an outside expert,” he said.
Corbitt previously created Emberall, a family history firm, before joining Y Combinator as an engineer and leader of firm School, the accelerator’s founder community.

He co-founded OpenPipe with his brother, David Corbitt, a former programmer at Qualtrics and Palantir who also co-founded a video legacy firm, GenerationalStory.

Read More: