Yupp

Yupp.ai operates as a platform for artificial intelligence model evaluation, distinguishing itself by integrating a crypto-based incentive structure. The company was co-founded by CEO Pankaj Gupta and AI lead Gilad Mishne in June 2024, both of whom bring experience from technology giants like Google and Coinbase. In June 2025, Yupp emerged from a six-month stealth period, announcing a significant $33 million seed funding round led by a16z crypto. The round attracted over 45 investors, including prominent figures like Google's Chief Scientist Jeff Dean, Twitter co-founder Biz Stone, and Pinterest co-founder Evan Sharp, alongside firms such as Coinbase Ventures.

The core of Yupp's business is a free-to-use platform that allows users to compare responses from a multitude of AI models side-by-side. It features over 500 models, including well-known names like ChatGPT, Claude, Gemini, and Grok, as well as lesser-known text and image models. Users input prompts and receive responses from two or more different AIs. The platform's business model hinges on collecting user feedback; users are incentivized to rate the responses and explain their preferences. This anonymized and aggregated feedback data, referred to as "preference packets," is then sold to AI companies to help them refine their models. This process is designed to address the challenge of AI model evaluation by providing large-scale, real-world human input.

To encourage user participation, Yupp employs a rewards system. Users earn credits for providing feedback, which can be used to continue accessing premium AI models for free. These credits can also be converted into fiat currency (USD, Euro, and over 20 others) through payment providers like Stripe and PayPal, or into cryptocurrency stablecoins on blockchains such as Base and Solana. This use of blockchain is intended to bring transparency and an immutable record to the AI evaluation process. The company has also introduced a public leaderboard called the “Yupp VIBE (Vibe Intelligence BEnchmark) Score,” which ranks AI models based on aggregated user feedback.

Keywords: AI model evaluation, crypto incentives, blockchain AI, human feedback, large language models, AI comparison, Pankaj Gupta, Gilad Mishne, a16z crypto, user-generated data, AI feedback platform, decentralized AI, preference data, model benchmarking, AI rewards, crowdsourced AI training, AI analytics, text generation, image generation, multi-model AI