{"slug": "intelligence-the-designarena-maker-raises-7-9-million-for-human-ranked-ai-evals", "title": "Intelligence, the DesignArena maker, raises $7.9 million for human-ranked AI evals", "summary": "Intelligence, the San Francisco-based maker of DesignArena, raised a $7.9 million seed round led by Index Ventures, according to co-founder and CEO Grace Li. The startup, which uses human rankings from its free product to evaluate AI-generated creative work, reported $60 million in annual recurring revenue, up from $5 million six months earlier, and claims 5.5 million users across more than 190 countries. The funding will support Intelligence's efforts to sell preference data to AI labs through private evaluations.", "body_md": "[Grace Li (@grx_xce)](https://x.com/grx_xce), co-founder and CEO of Intelligence, [said in a thread on X](https://x.com/grx_xce/status/2084361692792934488) on August 3rd that the maker of DesignArena raised a $7.9 million seed round led by Index Ventures.\n\n[https://x.com/grx_xce/status/2084361692792934488](https://x.com/grx_xce/status/2084361692792934488)\n\nThe San Francisco startup is trying to build a business around a stubborn problem for AI labs: software can test whether generated code runs, while judging whether a website looks good or a game is fun still requires people. Intelligence collects those judgments through [DesignArena](https://www.designarena.ai/), a free product where users compare anonymous outputs from competing AI models.\n\nLi arrived at the idea shortly before graduating from Harvard in 2025, according to [TechCrunch's report on the round](https://techcrunch.com/2026/08/03/designarena-creators-raise-7-9-million-to-bring-taste-to-ai-models/). Li and friends were building an AI game engine whose models produced functional games that were not enjoyable to play. The experience pushed them toward measuring subjective human preferences at scale.\n\nLi studied computer science and neuroscience at Harvard and previously worked at Apple. Her co-founder, [Kamryn Ohly](https://www.linkedin.com/in/kamrynohly), studied computer science and education at Harvard and also worked at Apple, according to [Y Combinator's profile of the Summer 2025 startup](https://www.ycombinator.com/companies/design-arena). Intelligence grew from a three-person founding team listed by YC to 10 people, according to Li.\n\n### Turning consumer prompts into evaluation data\n\nDesignArena lets users request websites, mobile apps, games, slides, images, videos, logos, audio and other creative work. It generates outputs using multiple models and asks users to rank them through side-by-side comparisons. Those votes populate public leaderboards and give Intelligence preference data that can be sold to AI labs through private evaluations.\n\nThe consumer product doubles as the data collection layer. Users receive access to a range of models without paying, while their choices reveal which outputs people prefer across formats, regions and time periods. Intelligence's site lists separate leaderboards for web development, mobile development, games, images, video, audio and slides.\n\nLi said Intelligence reached $60 million in annual recurring revenue, up from $5 million six months earlier. The twelvefold increase is self-reported. TechCrunch attributed the current revenue figure to Li and reported that Intelligence had closed an early deal with a frontier AI lab after demonstrating the need for design-focused human feedback.\n\nThe user count varies across Intelligence's public materials. Li's announcement said 5.5 million users across more than 190 countries, TechCrunch reported 5.3 million, and Intelligence's [careers page](https://www.intelligence.ai/careers) says more than 5.6 million. YC's original launch materials said DesignArena reached more than 47,000 users in 136 countries during its first four weeks. The figures point to rapid consumer distribution, though they are company-supplied counts rather than independently audited usage data.\n\n[Index Ventures (@IndexVentures)](https://x.com/IndexVentures) led the seed financing. [Conviction](https://x.com/conviction), including Sarah Guo and Mike Vernal, [A*](https://x.com/A_StarVC), Valkyrie and [Y Combinator](https://x.com/combinator) also participated, according to Li and TechCrunch.\n\n### A crowded evaluation market\n\nInvestors have already placed larger bets on the model-evaluation category. [LMArena raised $150 million at a $1.7 billion post-money valuation](https://techcrunch.com/2026/01/06/lmarena-lands-1-7b-valuation-four-months-after-launching-its-product/) in January 2026. LMArena also turns side-by-side user votes into model rankings, with coverage spanning text, image generation, web development and other tasks.\n\nIntelligence is narrowing its pitch around creative output and taste. That focus gives DesignArena a distinct data set, but it leaves Intelligence exposed to AI labs building internal evaluation systems or sending the same models to larger general-purpose arenas.\n\nCrowdsourced evaluation has also produced a visible failure. Yupp shut down after raising $33 million and saying it had reached 1.3 million users, TechCrunch reported. Its closure showed that a large voting audience does not by itself produce a durable evaluation business.\n\nIntelligence's reported $60 million ARR would put it on a different financial trajectory, if the figure holds. The seed round gives Li and Ohly capital to add engineers and researchers while expanding beyond a single leaderboard into a broader evaluation lab. Intelligence already lists experiments for predicting events, measuring social influence and evaluating conversations alongside DesignArena.\n\nThe underlying bet is that model intelligence is becoming abundant faster than reliable ways to measure it. As labs converge on coding and reasoning benchmarks, product quality will increasingly depend on subjective questions about aesthetics, usability and creative value. Intelligence is building the voting system that it wants those labs to use as the answer.", "url": "https://wpnews.pro/news/intelligence-the-designarena-maker-raises-7-9-million-for-human-ranked-ai-evals", "canonical_source": "https://runtimewire.com/article/intelligence-designarena-raises-7-9-million-ai-evals", "published_at": "2026-08-03 19:56:54+00:00", "updated_at": "2026-08-03 20:26:55.684514+00:00", "lang": "en", "topics": ["ai-startups", "ai-products"], "entities": ["Intelligence", "DesignArena", "Grace Li", "Index Ventures", "Conviction", "A*", "Valkyrie", "Y Combinator"], "alternates": {"html": "https://wpnews.pro/news/intelligence-the-designarena-maker-raises-7-9-million-for-human-ranked-ai-evals", "markdown": "https://wpnews.pro/news/intelligence-the-designarena-maker-raises-7-9-million-for-human-ranked-ai-evals.md", "text": "https://wpnews.pro/news/intelligence-the-designarena-maker-raises-7-9-million-for-human-ranked-ai-evals.txt", "jsonld": "https://wpnews.pro/news/intelligence-the-designarena-maker-raises-7-9-million-for-human-ranked-ai-evals.jsonld"}}