LinkedIn's New Crosscheck Feature Lets Premium Users Test Top AI Models for Free

LinkedIn has announced a new way for users to try out some of the latest AI models without worrying about token limits or paying for separate subscriptions. The feature is currently rolling out to LinkedIn Premium users in the United States.

LinkedIn lets users compare AI models with new Crosscheck feature

Per Hari Srinivasan’s announcement LinkedIn post, the feature is called Crosscheck, and it works like a side-by-side comparison tool for AI responses. Users can enter a prompt and receive two different answers, each generated by a separate AI model. Interestingly, the model names are hidden at first.

It’s worth noting that only after selecting the preferred response does LinkedIn reveal which model generated each answer. The company describes this as a “blind taste test” approach, which would help users evaluate outputs without bias.

Crosscheck already appears to support a wide range of models. In early testing, responses were generated using models from OpenAI, Anthropic, Google, Microsoft, as well as others like Mistral and Amazon.

That being said, the feature is still limited in some ways. Crosscheck only supports text-based prompts for now, meaning users can’t generate images, upload files, or access more advanced tools available on native platforms. However, there are no limits on the number of text queries, which removes the need to track usage or tokens.

LinkedIn also noted that anonymized data from these interactions will be shared with model providers to help improve performance across different industries. The company says no personally identifiable information will be shared. As of now, Crosscheck is still labeled as an early experiment from LinkedIn Labs.