Yupp, officially known as Ber Sarai Labs Inc., launched today with $33 million seed funding led by a16z crypto, the cryptocurrency-focused arm of Andreessen Horowitz, to build an artificial intelligence model evaluation platform.

The core idea behind the platform is simple: Instead of having a single AI respond to a prompt, users will get two responses — or more. This can be good if they want to see more than one opinion or avoid bad answers. It can also allow users to build a consensus with multiple AI models.

The company says that it has more than 500 models under the hood. This includes access to OpenAI’s ChatGPT, Anthropic PBC’s Claude, Google LLC’s Gemini and DeepSeek. It also features a wide range of open-source models, including Meta Platform Inc.’s Llama, as well as various text and image models from lesser-known research labs and small companies.

The reason Yupp is doing this? It’s because real human feedback and evaluation is necessary to accelerate best-in-breed development for AI models, co-founder and Chief Executive Pankaj Gupta said in the announcement.

“Your feedback helps tailor the AIs better for you, but it also helps the AI community build better models for everyone,” explained Gupta. “Feedback from one user may or may not be reliable, but aggregating data from millions of users around the world creates powerful signals that can be used by AI model builders to improve their systems and agents.”

As users browse answers from different AI models, feedback will allow users to share which one they liked and why. This feedback will help tailor future responses on Yupp to their needs.

To provide an incentive for users to stay on the platform, the company will offer users a type of token called Yupp credits. These credits go toward using the AI models. Users are rewarded these credits for providing feedback: The higher the quality of the feedback, the more credits. The token can be exchanged for other cryptocurrency payouts or cashed out for traditional currency.

“Beyond creating a better user experience, Yupp was founded to address a foundational challenge in the field of AI,” added Gupta. “Every AI builder wants to know how good their models or systems are, and for which use cases.”

To help developers handle these challenges, Yupp has launched a beta version of its AI leaderboard today. It will rank AI models by what the company called the Yupp VIBE leaderboard, or Vibe Intelligence Benchmark score.

“We believe crypto can bring transparency and ownership to this murky area of AI,” Chris Dixon, founder and managing partner at a16z, and Elizabeth Harkavy investor at at a16z, said in a statement about the investment.

Blockchain technology can make it easier for people to receive rewards for contributions. At the same time, companies have used them to provide AI builders with provenance for feedback data and evaluations. Yupp’s technology crowdsources assessment and uses the rewards to bring in users, but it also uses blockchain technology to provide transparency and an underlying immutable record of the evaluations that cannot be tampered with. AI needs strong, reliable evaluation based on large-scale human input, a16z crypto said.

In addition to a16z crypto, more than 45 individual angels and small investors joined the round, including Jeff Dean, chief scientist at Google; Biz Stone, co-founder of Twitter; Evan Sharp, co-founder of Pinterest; Aravind Srinivas, CEO of Perplexity; Kunal Shah, CEO of Cred; and Coinbase Ventures.

 Image: SiliconANGLE/Microsoft Designer