LinkedIn launches Crosscheck feature letting Premium subscribers test AI models

LinkedIn has rolled out a new feature called Crosscheck that allows Premium subscribers in the United States to test AI models from multiple providers without token limits or additional subscriptions. The feature, described by LinkedIn Chief Product Officer Hari Srinivasan as a "blind taste test" for AI models, presents users with two responses from different AI systems for each prompt. Users select their preferred answer before seeing which models generated each response. Crosscheck currently supports models from Anthropic, Google, MoonshotAI, Mistral, Amazon, OpenAI, and Microsoft. The system is limited to text-based prompts and does not support image generation, file uploads, or other advanced tools available on native AI platforms. LinkedIn shares anonymized usage data with model builders to help them understand how their AI performs across different occupations, though no personally identifiable information is included in this data sharing. While currently available only to U.S. Premium subscribers, LinkedIn plans to expand Crosscheck to more countries and free users in the future. The company also plans to add a leaderboard tracking how various industries rate different AI models.