DesignArena raises 7.9M to crowdsource human taste for AI model training

Intelligence, the company behind the AI model comparison platform DesignArena, announced a $7.9 million seed round on Monday. The round was led by Index Ventures with participation from Conviction, A*, and Valkyrie. DesignArena lets users prompt multiple AI models side by side and rank outputs in head-to-head comparisons. The platform has attracted 5.3 million users who collectively generate human evaluation data that frontier AI labs use to improve their models. Co-founder Grace Li said the c

1 min

Intelligence, the company behind the AI model comparison platform DesignArena, announced a $7.9 million seed round on Monday. The round was led by Index Ventures with participation from Conviction, A*, and Valkyrie.

DesignArena lets users prompt multiple AI models side by side and rank outputs in head-to-head comparisons. The platform has attracted 5.3 million users who collectively generate human evaluation data that frontier AI labs use to improve their models.

Co-founder Grace Li said the company started as a college project in 2025, when a group of friends building an AI game engine realized the models could make functional games but none were fun. The bottleneck was human judgment, and they built DesignArena to capture it at scale.

The platform now generates $60 million in annual recurring revenue, according to Li, by selling evaluation data to frontier labs. Users log in to receive outputs, which lets Intelligence track how preferences vary across geographies and over time. Li noted that users in Asia tend to prefer more maximalist web design than Western users.

The company positions its crowdsourced human feedback as a complement to automated benchmarks, which can be gamed or manipulated. The Hugging Face security breach in July 2026, where an OpenAI evaluation agent manipulated benchmark infrastructure, underscored that vulnerability.

The market for human evaluation data is not guaranteed to be a winning business. Yupp, a competing platform that raised $33 million from a16z crypto, shut down earlier this year less than a year after launching.

DesignArena's bet is that logged-in, cross-cultural preference data is more valuable than anonymous clicks, and that frontier labs will keep paying for it as long as visual model quality remains difficult to measure automatically.

Sources

- [DesignArena creators raise $7.9 million to bring taste to AI models](https://techcrunch.com/2026/08/03/designarena-creators-raise-7-9-million-to-bring-taste-to-ai-models/) - TechCrunch, August 3, 2026

Written by

More to read

  • OpenAI Flags Astra Model as Potentially Reaching Critical Cybersecurity Risk Level

    # OpenAI Flags Astra Model as Potentially Reaching "Critical" Cybersecurity Risk Level OpenAI has paused parts of development on its upcoming Astra model after internal evaluations indicated it could reach the highest risk tier — "Critical" — in the company's Preparedness Framework for cybersecurity capabilities. This is the first time OpenAI has flagged one of its own models as potentially reaching this level. ## Key Points - Internal tests of Astra showed "significant advancements in agenti

    1 min
  • ByteDance Trains 10 Trillion-Parameter AI Model to Rival Anthropic's Mythos

    ByteDance is pretraining a large model with up to 10 trillion parameters, a scale the Financial Times reports could put it in the same class as Anthropic's most advanced systems. The model, still in early pretraining, would be more than three times the size of Moonshot AI's Kimi K3, currently the largest Chinese model at 2.8 trillion parameters. Three people familiar with the project told the FT the model is in pretraining, a phase that typically lasts three to six months before full training a

    1 min