Harmonic, the Robinhood CEO’s AI math startup, launches an AI chatbot app

Harmonic, an AI startup co-founded by Robinhood CEO Vlad Tenev, announced Monday the beta launch of an iOS and Android chatbot app where users can access its groundbreaking AI model, Aristotle. This strategic move aims to significantly broaden access to Aristotle, which Harmonic boldly claims offers “hallucination-free” answers for questions involving mathematical reasoning. This is a remarkable assertion, given the pervasive reliability challenges faced by today’s advanced AI models.

Harmonic’s core focus is on creating “mathematical superintelligence” (MSI). The startup envisions a future where Aristotle can assist users across all fields that heavily rely on mathematics, including complex subjects like physics, statistics, and computer science, thereby revolutionizing problem-solving in these domains.

“Aristotle is the first product available to people that does reasoning and formally verifies the output,” stated Harmonic CEO and co-founder Tudor Achim in an interview with TechCrunch. He further emphasized the company’s confidence, asserting, “Within the domains that Aristotle supports, which are quantitative reasoning domains, we actually do guarantee that there’s no hallucinations.” This commitment to accuracy sets Aristotle apart in the rapidly evolving AI landscape.

Looking ahead, Harmonic plans to expand Aristotle’s accessibility by releasing an API for enterprise clients, alongside a dedicated web application for broader consumer use, ensuring its powerful capabilities reach diverse user bases.

Harmonic demonstrated Aristotle’s prowess by achieving gold medal performance on the 2025 International Math Olympiad (IMO) through a formal test. This means the problems were translated into a machine-readable format, allowing for precise evaluation. While Google and OpenAI also developed AI models that achieved gold medal performance on this year’s IMO, their tests were informal, relying on natural language. Harmonic has stated it is not releasing other benchmarks for Aristotle at this time.

The beta launch of Aristotle follows closely on the heels of Harmonic’s impressive Series B funding round, which successfully raised $100 million. The round was led by Kleiner Perkins, valuing the startup at a significant $875 million. Achim expressed confidence that Harmonic is “advancing very rapidly along” its path to achieving MSI, a trajectory that investors evidently believed warranted such a valuation given the ambitious scope of the startup’s goals.

The pursuit of AI models capable of solving complex math problems is a key focus for several leading tech companies. While mathematical proficiency in AI is valuable in its own right, math is also considered a uniquely verifiable domain that demands core reasoning skills. Systems that master these capabilities are expected to prove immensely useful across a multitude of other domains, extending their impact far beyond pure mathematics.

Achim elaborated on how Harmonic achieves its hyper-accurate solutions: Aristotle produces responses in Lean, an open-source programming language. Crucially, before Aristotle delivers an answer to users, the model double-checks the solution’s correctness through an algorithmic verification process that does not involve additional AI. This methodology, Achim noted, mirrors technology used to verify outputs in high-stakes fields such as medical devices and aviation, underscoring its reliability.

Even within a narrow domain, achieving hallucination-free performance from an AI model remains an incredibly difficult challenge. Studies have consistently found that even leading AI models frequently hallucinate, and the issue does not appear to be diminishing. In fact, OpenAI’s latest AI reasoning models have shown to hallucinate more than their predecessors, highlighting the significance of Harmonic’s claims regarding Aristotle’s accuracy in mathematical reasoning.